Salesforce Stock

Please welcome to the stage, EVP, Global Investor Relations, Mark Murphy.

Thank you. Welcome, everyone. Great to be here. Look at the energy in this room. Love it. Love it. As mentioned, I am Mark Murphy, EVP of Global Investor Relations with Salesforce. We have just a fantastic Investor Day lined up for you. I want to start by thanking everyone in this room. We are so grateful to you for your time and attention. We hope you're going to make the most of this unbelievable event by getting out there and engaging with customers and partners, ask them anything you want, ask them what's changing, ask them how this transformation is occurring for them.

This is at least my 20th Dreamforce. If you can believe that, I've kind of lost count, but it's at least 20. And it's just unbelievable to be here and to be on this side of the room for the first time. It means so much to me. And as I reflect back on having been here through 20 Dreamforces, what is actually amazing is what has not changed at all. The location hasn't changed. We're still here in San Francisco and Moscone Center. The keynote is in the same room. The founders haven't changed. We have Marc, we have Parker sitting right over here. They're going to be on stage later today. The energy level hasn't changed.

At the beginning of the keynote, the Hawaiian ceremony hasn't changed. I think a lot of the jokes that get made have not changed somehow in 20 years. What has changed is the technology, and the technology has changed quite a bit over the summer. There's a new ability to unlock trapped value. That has legged up in a pretty exponential way in the last couple of months.

And so because of that, we're going to have a totally different structure to the Analyst Day. It's going to be unlike other Analyst Days that you have seen. We are going to be opening up laptops, and we're just going to be showing you live this kind of transformation that is happening. If we're going to be trying to show you what's possible, we're also going to have many more customers on stage. We're going to have Siemens. We're going to have Adecco, and we're actually going to have Anthropic on stage toward the end of this event.

One more thing that has not changed in the last 20 years is the legal disclaimer. And so I'm not going to read it to you, but we're going to make some forward-looking statements that are subject to change. And we would strongly encourage you to refer to our most recent SEC filings, including 10-Qs and 10-Ks. So this agenda is going to be fast and efficient. We're going to really fly through this. We're going to start with Patrick and Rohan. They're going to go a lot deeper on the technology.

We are then going to move to the go-to-market execution engine. We're going to have Miguel and Alexa for that section. And then batting cleanup, we're going to have Robin with the financial framework for this Agentic opportunity. After that, Marc is going to join us, and Marc will be on stage, and we will have a Q&A with the full leadership team for as long as they want to go at the end. And after that, we're going to have an investor reception and demos. And that's actually going to be right outside of this room.

So with that, I'm going to hand it over to Patrick, who is President of Applications and Marketing. Patrick, welcome.

All right. Thanks, everybody. Does anybody want to take any bets on whether this agenda is fast and efficient, as Mark put it? Because I'll be happy to take a couple of bets. I think I know where. But that's okay. Okay. So I'm psyched to be here with you all today. Don't look at me like that, Miguel. And I -- we're going to walk you through a whole bunch of stuff about where we're going. We're going to kind of break down what you saw in the keynote a little bit more today.

I have been getting a lot of questions from yesterday on how did we do that. There's -- it's funny, there's like 2 sides of belief. There's the side that kind of has an inkling and kind of wants to know more detail and then there's the other side that's like that couldn't have been real. And it totally was. So we're going to break that down and show you how all of that works.

Rohan and I, the way to kind of think about these 2 sections in the agenda here is I'm going to start by walking you through how I think our apps are going to change and how they need to become agent-ready or a more technical term agent legible, meaning every app needs to be usable by agents in a composable environment like Cowork like you saw earlier. And then Rohan is going to come up and start walking you through how we are in the midst of transforming into a much bigger data company and how we'll leverage that data in an enterprise harness for the future. So that's the course of the next 30 minutes or 40 minutes or so.

Now what I want to do is something just mildly annoying for a moment is I want to hop out of the slides, and I want to go to the demo screen just for a second, if we could. Thank you, team, back there. Okay. So I mentioned -- actually, how many of you saw the keynote yesterday? Oh, most of you. Okay. Great. I thought it'd be about 1/3 of you. Good job, everybody. So I mentioned in the keynote yesterday that you can go from nothing to something pretty usable in about 6 to 8 minutes. And then after that, you might have a few days of iteration as you build and you refine and you kind of design the user interface that you want.

So I want to try to bring that to life a little bit here. So I've got a prompt up here, and I worded it kind of in a funny way, but I'm saying build a one-shot artifact and artifact is actually a capability within Cowork, which you'll see in a moment. And I wanted to list my open Salesforce opportunities. I wanted to use the headless toolkit to do it. And then I wanted to build a model where Claude drafts a task and posts that over to Slack. And then I say no questions. One shot at the front and no questions at the back are me trying to get Claude not to ask me any damn questions and just build the thing so that I don't have to keep coming back in and hit Enter while this is going.

And what I'm going to do is this is like the cooking show. If you've ever seen a cooking show, where they pull the turkey out of the oven, even though it's been in there all day, is we're going to hit Enter, and you're going to see this start, and it's going to start working on this application, this opportunity application that I just asked. It's loading the tools. And all in all, this is probably going to take about 6 minutes end to end to come back with a completely working application.

And this is really -- this is no longer vibe coding, right? This is we're taking the acceleration that these coding agents have delivered to developers, and we're putting it in the hands of knowledge workers. That's really what Cowork is. There's -- you're never going to see any code here. You're not going to see any of that background. Coworker just takes care of all of it for us. In fact, we can kind of expand and see what's going on. Now this is going to take maybe 8 minutes or so. And so we're just going to let that run, and we're going to go back into our slides. And then when we come out of our slides, God willing, it will have produced something mildly useful, and we'll take a look at it.

Okay. We can go back to the slides. Okay. So I think you heard yesterday that we really believe that there is a big interface revolution going on. Another way that I like to say this is that there is a user behavior or a user experience revolution going on, which is that most people in the world in software, at least, are starting to figure out or starting to sense that there is a better way to use software, right? The way that we've used software for 50 years is we have to know what we want to do, and we have to know then how to use the software itself, and we have to click around and hunt and peck.

The analogy that I like to use is that if you're a digital photographer, like a wedding photographer, part of your job is being -- having a good eye and knowing how to use your camera, but like most of your job is just knowing how to use Photoshop, right? Like it's a big complicated application. Salesforce is the same way. It's a big application. It has a ton of surface area, and your ability to extract value out of it is limited to the human's ability of what they know about how to use the application.

And what's happening now is we are taking an agent and putting it between the software and the human. And we're just taking intent from the human. It's giving it to the agent, to the knowledge worker agent. In this case, what you'll see today, that knowledge worker agent is Claude. Claude is looking across all of the tools and capabilities, which have been connected inside of Cowork, and it is composing a response based on all of those tools. So you are getting a tremendous amount of additional value out of not just Salesforce but any other platforms that you plug into it.

The other interesting thing that's happening there is it's kind of aggregating software into one environment because no longer do I have to do a discrete task in Salesforce and then a discrete task in SAP and a discrete task in Workday, I can -- if I have a task that requires all 3 of those systems, if they're all wired up in the same way that we just showed you yesterday like we can do with Salesforce, if they're all in that Cowork environment, it's now all aggregated and that agent can go off and kind of work and operate across those systems in ways that, frankly, we've been trying to build integrations for, for the last 50 years. And now we have this entirely new way to kind of drive these integrations. So it's a big, big revolution. And of course, we've been working towards this.

Marc talked through this yesterday, really trying to simplify our platform to make it ready for this user behavior change that we see happening. We've really focused on our deterministic systems, of course, our data layer, which Rohan will go into; our apps layer, where we have a lot of the semantics and business processes that are codified into Salesforce. And then we added our Agentic layer and then now the new AI Force layer on top. And that's really where all of this starts to come together.

So AI Force, you can think about it as 3 kind of new surfaces. One is Claude Force, the second is Slack, the third inside of Lightning is Agentforce Coworker. We will do more of these in the future. There is no shortage of knowledge worker apps that are beginning to emerge where you would want to be able to use Salesforce in the way that I'm about to show you. And then, of course, all of that comes together on this platform with Data 360, Customer 360, Agentforce and AI Force.

When we click into Claude Force just for a moment, and I'll show you this in the app. Basically, we came off of Headless at our developer conference back in March. We launched Headless. We kind of -- we sensed the user behavior change that was happening. And we were like, we better get MCP servers out because our customers are trying to do this themselves, and it's a little bit dangerous. So let's get some first-party MCP servers out. That was, I think, very successful and a little surprising.

I don't think people expected that posture from a software company to actually endorse that pattern. We got out there with it quickly, and then we started seeing people pick up and use it very quickly. And then on top of that, we started seeing them run into problems that we needed to fix. And Claude Force is really the extension of that and AI Force in general. It is us productizing the capability of those MCP servers into something that can get implemented and deployed across your organization way faster than every individual in the room having to connect their own MCP servers, which is just really not very practical.

And then, of course, we'll talk more about this later across the sections, but then we wanted to package all of that up into one relatively easy to buy platform addition. So we have our new AI Force Max edition at $550 per user per month, and it includes everything that you're about to see today, including Slack and Slackbot, which you'll see glimpses of for me today, but not a ton.

And despite this new addition, for the time being, Claudeforce is in open beta. So we're trying to maximize test area of customers that can touch it right now. One day soon, that will move into a GA, and you'll have to buy it to keep it turned on. But right now, we really want people to touch it.

Okay. So I'm going to go back to the demo, and I'm going to break down a little bit what all that looks like. And then I actually have a customer -- a little bit of a surprise for you in a moment. We're going to bring a customer up, and we're going to see their live Claudeforce environment against their live production data as well. So let's go back to the demo. And you'll see -- there we go, I have to learn how to scroll. So it's still -- it did ask me a question. It's Claude managed, just constantly asking me a question. I'm just going to say, roll up command. I'm not really reading what it's asking me, so God knows what we're going to get. But I do have some confidence that it will be pretty good.

Okay. So we're going to let that keep running, and I can kind of go into into the rest of the demo. So you saw this big beautiful dashboard yesterday, right, and all this whiz-bang stuff happening on the screen. But that's not -- that's a capability within Cowork. It's an incredibly exciting one, but it's also not the way most people start using Cowork. The way most people start using Cowork is they start on the chat screen. And so that's exactly what I'll do. I'm just going to create a new chat here, and they just start asking questions. Before anybody figures out that they can build applications, this is the way people usually start.

And so what would a typical Salesforce question look like? Well, it might be, tell me about my pipeline, say, open pipeline. So a fairly typical question. Now what's going to go on in the background here as this thinks, and you'll see the work going on here is, first, it looks at the question and it's deriving the intent from the question. So just from the word pipeline, it is looking then at its tools that have been connected, and it is trying to connect a tool to the intent, okay? And in this case, it's going to connect Salesforce to that intent. It's going to say this is a question that Salesforce can likely answer, and then it's going to go ask Salesforce that question through the MCP server, and then it's going to bring back an answer. That is what is happening behind the screens.

And not only did it bring back an answer, but you'll notice that it brought it back as a dynamically generated kind of mini application right here in the channel. And that is all made possible through a new protocol within MCP called MCP apps. Sorry to get mildly technical for a second. You'll also hear us refer to it as HXL, which is just our internal name for it. And all it means is when we ask that question, Salesforce on the other side, it says, not only do I know the answer to that question, but let me give it back to you with some additional detail about what an application could look like on the other side. And then Claude says, thank you very much for that information, and it draws the application right here in the chat. And all of that took place everything that I just said took place in -- well, way faster than I was even able to tell you what was going on, right? So it came back with our answer.

And what's cool about these little mini applications is they're actually somewhat interactive. I can scroll over them. These aren't just like images that come back. I can interact with these things as well. In fact, I can even design buttons in. And all of this kind of sits on the Salesforce side. We send these instructions over to Claude and then Claude puts it on the screen. So I've got my North Star app here. This is an opportunity. This is actually the one that we were looking at yesterday at $4.8 million opportunity.

And let's say, I wanted to ask a follow-up question. So tell me about the open activities. And effectively, you're going to see the same kind of thing happen. It's going to go to Salesforce or first it's to read the intent. It's going to realize that this is a Salesforce question. It's going to look up the skills and the tools to see if it knows how to go get it. And then eventually, it's going to come back with an answer. And ideally, that answer is also going to be some sort of dynamically generated interface, which I expect that it will be as soon as you see Headless toolkit, yes, there we go. So these are all of the activities that are logged inside of Salesforce, just brought back immediately to me and even a little activity chart, kind of a heat map where you can see when the activities were happening.

Now these seem like relatively simple queries and they kind of are, but I want you to think through how much kind of internal human clicking is going on inside of Salesforce to do this the old way, right? You have to know how to get to opportunities, you have to load all of your opportunities, you have to click into every single one. You have to look at the entire opportunity screen for every single opportunity. You have to commit to memory, what's going on with those opportunities, then you have to synthesize that as a human being, and then that tells you what you should go work on or you can just ask Claude. And that's effectively what we've done here with Claudeforce.

Now let me show you a couple of other things. One thing that's very cool is, if there are certain things that you do every single day or every few hours or whatever it is, you can schedule these. So that's what I have. I have a daily briefing skill here. So here's the daily briefing. So this tells me what I should be worried about every single morning. And I don't even have to come in and run this every single morning. I can just tell Claude to schedule it. It saves it as a daily briefing schedule, which you see right up here.

And in a funny way, what you've done there is you've basically created -- just by asking, you've created an agent. You have created an agent that every morning is going to go out. It's going to inspect Salesforce. It's going to build your daily briefing, and it's going to bring it back and drop it right here in Claude. And of course, that goes to your phone as well. You get a notification every morning. So it's an incredibly powerful capability.

I can also show you -- so here is the command center that we looked at yesterday, see if the sound is actually working. It is, which is unfortunate for Parker because that means Parker is going to come up. There he is. And we're just going to minimize Parker because I think he's heard enough of that. But this is a very fancy version of that prompt that I started with, right? Remember that first prompt where I was like build me an application. That's where this started. And then it just grew and grew and grew in capability as I asked it for more and more, which I think somewhere -- yes, right here, you can kind of see all of those prompts going through, and there's actually more than this behind the scenes.

In fact, I can even show you. So remember a few moments ago, I told you that the kind of cool thing about this is the aggregation of software. This isn't just Salesforce, it's Salesforce and anybody else who builds these kinds of MCP connectors. Well, one of the companies that's really out in front on that is this company -- I don't know where I put it, here it is, is this company called ElevenLabs. And everything you see here on this screen, as I scroll through it, is if I know how to scroll, I actually don't even know how to scroll on this, I guess, we'll go this way. There we go. Everything you see here, I did not create. This is ElevenLabs internal kind of capability for creating voice models and for creating 3D avatar models. I didn't touch this. I didn't create any of this.

All I did was create an ElevenLabs account, hook up the MCP server. And then when I was building my command center, I said, here is a graphic of Parker in the silly lightning costume. And then I gave it an audio file from a keynote that Parker has done in the past. Those were the only 2 inputs. And the Claude connected to the ElevenLabs MCP server, did every single thing else. Really powerful. And actually, there was a funny moment in here, if I can find it. It might be here. I forget, yes. I look at all the silliness as it generates this all out.

I think it's here, filtered. Yes. So what happened is ElevenLabs didn't like this graphic because it thought that it demonstrated violence because of the lightning bowl, and it actually stopped me and prevented me from going further. And then we just had to change the initial graphic to make it look less like a real spear, and that's how we got through this. These AIs are constantly stopping you to make you do things. But anyway, none of this was -- 0% of this was manually created. This was all just from those series of prompts and the capabilities all hooked up.

Okay. So I'm going to wrap up here quickly by going back to -- if I can find it, where is this? Probably in here somewhere. Is it this one? No? Stephen, help me out here. The very top, this one? Keep going. Here, it raised its hand. Okay, that's a good sign. That means it's about done. Okay. So it's ready to publish the artifact. The artifact is the application that I asked it to create at the beginning. So if I accept this, boom, there is my pipeline, and I should be able to expand this and maybe shrink this a little bit.

So there is my full roll-up of all of my opportunities across my team. Super simple, just from that one silly prompt at the beginning. And I can make this better and better. I can make it look like whatever I want. For example, there was -- there's mine, which looked super fancy and had all the whiz-bang stuff, but what if I wanted it to look like a newspaper, all you'd have to do is ask in the instructions and you get a totally different view of all of your Salesforce data.

What if you wanted to build something that looks more like a deck. So this is Salesforce information kind of in slide format. So you could actually build a full presentation. If you had to go do a Board presentation, everything is just there, all from a couple of sets of instructions. And then finally, what makes this all work? Well, it's pretty straightforward. It is this new plug-in behind the scenes where we've taken the many, many MCP servers that make Salesforce work and a handful of skills.

Right now, most of those skills are focused on sales questions and salespeople, but we'll be expanding across our entire platform. And then things just kind of start to work because you have this powerful agent sitting between the human beings. So the key here is like this is not just about getting answers to your questions. This is about building full form applications and dashboards that you need to go get your business done. In fact, there was a thing that happened this morning. I'm not going to bring it up, but there was a -- I am going to kind of bring it up.

There was a thing that happened this morning that I woke up and the very first thing that I said is I need a command center, so I can see the surface area of the problem that's going on, and I just jumped into Claudeforce and built it out and then got my team rolling on it. So we're using it in real life. This is rolled out to Miguel and Alexa's entire organization. It's rolled out to the Anthropic selling organization. We have about 40 customers in the beta -- well, in the pilot. As of yesterday, when we put the beta link up, I actually haven't checked to see how many customers have signed up for the beta, but it's probably several hundred or 1,000.

And with that, I would like to bring up one of those customers. We're going to bring up Rick Janssen, who is the CIO of Siemens. You heard a little bit from Siemens, and we'll invite Rick up, and he's going to show us how he is using Claudeforce.

We didn't test his laptop before that. So we'll see.

Hi everybody. Thanks for having me.

Thank you. I appreciate it.

So let's see whether this is working without any [indiscernible].

So this is going to be fun. We turned Claudeforce on for Siemens, what, 1 week ago, Frederik, 2 weeks ago.

There we go. Very good.

When do we turn this on for you?

A week ago?

A week ago.

A week ago, roughly.

Yes. And you've built the full sexy command center with a 3D avatar of yourself by now as well, right?

I actually had my team build a little avatar view, which I wanted to show up here, but...

You know what, I kind of deserve it actually, if you did that. Why don't you show us where you are?

Yes. Okay. So I mean, first of all, to start with, we are really excited. And I would fully agree from a Siemens perspective to what you have been calling the interface revolution. We believe that this is really the necessary step. And it's great to see that Salesforce is taking that, some might say, disruptive step to open up. I think the architectural readiness with Data 360, Headless 360 is there, and now introducing the partnership with Anthropic is just a great additional step. And I mean, as you can see on my machine, I'm running Claude Desktop. And to also quickly share here in the settings, we are running Salesforce for sales, the plug-in, right, what you just shared.

I also have the team provide me 2 skills. 1 is the Siemens report, so just a layout and how we would like to create reports. And then the other 1 is the Salesforce data masking. We are, of course, looking at productive data here. So we want some data orchestration for the report we are generating in a second tab.

We were coming up with this on Monday night. We were sitting in the W with his laptop thinking about what to prompt. And he's like, but the problem is it's going to send production data. Miguel was trying to figure this out as well for his demo. And I said, no problem, just tell Claude before it outputs the answer to obfuscate your data, and it will do it. And it does. So we'll take a look.

So, and I mean, what I prepared for the demo is basically a prompt where I want Claudeforce to list me the top 10 Dutch accounts and their opportunities. And I want it generated as a report based on the skill I was just sharing in the skills section. And I think let me now hit return. And of course, we're going to have to bridge a little bit of time here just starting to think and execute.

And the fun thing about these demos is the AI works for a little while. And so you have to riff in between. The voice demos are also challenging because you don't mute in time, then they don't, it's like a whole new world of figuring out how to do these demos. How is that? Terrific? Did I go long enough?

No.

No.

No. No. Maybe we can quickly discuss where we could take it next, right?

Well, before, actually, maybe before where we go next, so this is, you said, a little bold for Salesforce to do, but maybe equally as bold for you, right? I mean this is an AI that now, click again, I told you that AI is constantly asking you. This is an AI that is like looking at and reading your Salesforce data and in some cases, maybe even writing your Salesforce data, maybe a scary thing, but you jumped into it. Why?

A, because we believe, as I just said, the revolution has just started and maybe you're delivering the blueprint for what many other companies would follow. You maybe saw our CEO also on stage at the keynote. And I think for him as well, it triggered somewhat, some kind of process that he says we need to maybe open up in the same way. I mean you're the #1 in CRM. So you have all the data about the customer and about how you are serving the customer. We have all the data about the product. So you have the customer truth, we have the product truth.

Absolutely.

So basically, similar path. Another thing is that, I mean, this is what I would call the Claudeforce version 1, which we're now launching. It takes away the need for many of our sellers, people in service and marketing, to really go through the already established interfaces, which we know from the past. But it would most likely start with reading data, combining data.

So you're starting with read... and synthesis...

Correct. But of course, the real magic would happen as soon as we would be able to also write in a meaningful way back into the systems because that is then doing exactly what you also were pitching. It takes away the elements where software has the human do all the work, and that is, I think, the next level of productivity, which we, of course, also want to achieve.

Absolutely.

And there's maybe even cool things we can do together.

We should have started this 1 before, just like I did the other prompts, but it's rocking and rolling here. This stuff is going to get...

Still going...

Still going. This stuff is going to get way faster over time. Really, what's happening in the background, it's not latency within the platforms. What it is, is overthinking is what it is. It is over, right now, the AI is over reasoning over the tools that it has available to it. And so all of that can be tuned as we move further in. And this is exactly why we want to get this out to as many customers as possible so that we can identify where you have what should be a simple question being over reasoned on as you come back in.

Another way to solve that also would be to change the model and go to something like Sonnet, which will do much less reasoning and thinking, and it will just kind of immediately, the bigger the model you use, the more thinking it's going to do, which, by the way, is also more tokens. So that's relevant as well. Let's see, can you, let's see if we can expand, gives you a little expand on that. It might not. Yes. So this is what it's doing. You can kind of see all of the thinking. Actually, this is what's called the chain of thought. You're actually seeing step by step what the AI did there. So I got the Dutch accounts, now I need to. This is like literally, it is writing down its thinking as it goes, and this is incredibly useful information for us from a tuning and performance perspective as well if we have the telemetry to look at it.

So it should be ready in 1 second, but I think it's also worthwhile mentioning the additional connectors, which we would be able to bring in overtime. Yes. So connecting to the Microsoft Office environment, having all the Office 365 topics like conversations with customers, e-mails, documents, Exchange tasks noted down, documents are shared in SharePoint, et cetera, bringing that in as additional context, of course, something which would also be a very meaningful addition. I think the teams are working on making that already happen. And then the SAP connection is also something which is kind of powerful because we do billing, cash collection, everything on the SAP side. So having the end-to-end view on what's really happening in the different customer accounts.

How much of that from the Salesforce side, we've kind of solved a lot of that for you. You said teams working on it. For the other platforms, maybe not so much yet right? You guys are wiring it up yourself.

Right. We are wiring up it ourselves, but I would expect that you're going to have connectors, which are maybe then double tested or double checked with the different other tech partners to really have it rock solid and then out of the box basically.

Yes. Yes. I got you. So, this is what I meant when I said fast and efficient isn't going to work out. We're still rocking and rolling here.

I did somehow prepared for that scenario and now...

Yes...

I've created... perfect. And I maybe just open it here in Google Chrome so that everybody can see it. I want a nice extension, but I think here we go. And those are now the top 10 Dutch accounts, right? Obviously, it is, as I said before. But you would find it by value. In the Siemens report, as we define it in the skill upfront, you could drill down into the opportunities. So you see where your pipeline sits, what the status of the different deals is, and which stage they are, whether they are closed. And then also what patterns maybe stand out, win rate, cancellations. And then you could take it, of course, and dive deeper and double-click on the different elements to better understand what action you might want to take or to identify what maybe other people did to fix situations like this in other customer scenarios.

You'll notice it did something which I didn't expect, which is it actually went and got your real logo, and it probably even went to your website and stole some of your CSS to kind of make it roughly look a little...

That's the skill I showed...

You built in the background.

Yes, exactly because then we know that the system would always provide it in that kind of style and layout.

Incredible. Well, where are you going next with this?

As I said, I think we want to make sure that Claudeforce is really used across the organization. We have now the pilot going on with 25 users. The first initial feedback is excitement. People love it. And so I think we want to sit down and need to figure out what we are agreeing in terms of the pricing model behind it. But ideally, we can bring it out to our sellers ASAP and then innovate from there because I think that is what now the market is expecting. And I think we can speed up the transformation of our company with your support.

You have 25 now. How many do you think it could be?

18,000.

There you go on the go. 18,000. Amazing. Amazing. All right. Well, Frederik, really appreciate it. Thanks so much. This is super exciting. If you have any problems, then you will, you'll run into a few. Yes, please go to me, and we'll get it picked up, and we'll chase those 18,000 users.

Perfect. Thank you.

Thank you so much. All right. So I think from here, we're going to shoot it over to my colleague, Rohan, and we're going to start talking about trusted context and data. And Rohan, over to you. Thanks.

All right. Well, it's fantastic to be here and be with all of you. I, Rohan Kumar, I joined Salesforce 3 months ago, and it's just been a lot of fun. As Patrick started off, there's 2 big focus areas. Obviously, how do we get all the value that we have to be agent readable. That's a huge focus with headless and the work we are doing there. And the other big piece that I'll get into is where we are transforming to become a data company. And I'll talk about why is that and the importance of why we need to go there. And it's interesting, you look at like a lot of the value that AI is accruing, there's 2 very important things for these agents to be very successful. It's the model itself, which has the intelligence. It knows a lot about the world.

The challenge is the model doesn't know anything about the business, right? When you look at customers, products, employees and how the interactions happen, the history and memory of how your business functions, a lot of that, the model doesn't understand. So by itself, no matter how intelligent the model is, it can't reliably reason and act on behalf of your company. And that's really where the data and the enterprise context to like that understanding of your business becomes very important. The challenge is building the enterprise context is very hard. There's, you look at like the data is in silos. A lot of times, the decisions that get made, they get made in meetings, there's no structure we are capturing all of that. And without having that comprehensive understanding, this is where agents hallucinate, right?

When you look at how do you build trust inside agents to do complex stuff, getting that enterprise context right becomes very critical. And what I'll walk you through is how do you go create that? And it's, creation is 1 thing, but this is the most important IP of a company. And with these agents, you got to make sure that you secure it and then you have to manage it at scale when you have like a huge digital workforce that's running in your enterprise. So the first step in terms of creating the enterprise context is you have to get your data ready for AI. And what I mean by that is today, if you look at any reasonably complex enterprise, they have many SaaS applications, which have very high-value data. They essentially have data lakes, data warehouses, storage systems, where a lot of high-value data is getting stored.

And it's actually really hard to even discover this, right? And so we have this product suite called Informatica that essentially helps data leaders in an enterprise discover all their data assets, clean them, connect them, figure out the semantic meaning and make that, take that first step of having the data ready for AI. So before the agents really understand your business, it's important that our customers understand the data assets in the first place. So that's step number 1. Now once your data is ready, it's ready, the next step essentially is to create what's called the enterprise context. And there's a difference. Data being ready means you can have like a bunch of structured data in your SQL tables that needs to be synthesized into figuring out what does revenue mean for the company, right?

So that's a higher-level concept as an example, or what customer segmentation means for you. And to sort of create that context, we have this product called Data 360, which essentially not just relies on Salesforce data because there's obviously a lot of your enterprise data that's outside of Salesforce. And we've built this cool technology called Zero Copy. And what that does is it's able to reach into all the data systems that have been discovered by Informatica, remember, the first step. And then without physically copying the data, it looks at the metadata of each of these systems and it's able to synthesize the right context in a central location. So that's the second step. So you've made the data ready for AI with Informatica. And then Data 360 using the Zero Copy is able to get your Salesforce data and everything else into your context layer.

Now once you have built out the enterprise context, the third step is essentially creating a semantic model. And semantic model essentially is your business glossary. When you think about what does customer churn mean, what does ARR mean, what does customer health mean? These are not just some static fields that you see inside a database. These are like business metrics, business relationships that have been created over decades, which are very important, like this sort of information for an agent plans are becoming very important. So Tableau is the tool that we have, the product suite that we have that helps you create these semantic models. And it's interesting, these manifest as dashboards today. If you look at Tableau is the BI tool, it creates dashboards. And of course, dashboards is being commoditized right now, like every AI tool helps you create dashboards.

But the value of Tableau is never the dashboard itself. It is the semantic model layer that gets manifested at the dashboard. So in fact, if you think about, if creations of the dashboards become very common, you can do that in any AI tool, the value of your semantic models essentially has really gone up because if your semantic model is not right, then all those dashboards are actually giving you false information. So that's really where Tableau comes in. And now you have all this trusted context plus semantic model that you've created, it's available to all your agents. I mean this is great because it makes the agents a lot more accurate because they're working on a deep understanding of the enterprise. And then it also makes them more efficient. So they're not burning a lot of tokens.

I think the point that Patrick was making where if you don't have the right data organized, then the agent is spending a lot of time trying to get to the right information, which is where they're spending a lot of cycles and a lot of tokens. So it becomes very expensive. So creating this enterprise context early on lands up becoming very important. The challenge, of course, is once you have all this information, if you put yourself in the shoes of the security leader, it's extremely risky because if these agents that are using this context start leaking it, deleting it, that can become a huge problem. So this is where we introduced this product called Salesforce Guardian. It's the evolution of what we have with trusted services in Shield. In Salesforce, we basically had capabilities like encryption of your data, data masking, et cetera, to protect your assets, and that's evolving into 2 additional capabilities.

1 is called agent identity, which is the embodiment of how the agent runs. For very complex stuff, you don't want the agents to run on behalf of the user because that gives them a lot of privilege. You want them to have identity which can be controlled. So that's 1. And the second thing is actually protecting this enterprise context itself, like capabilities that can actually classify your data, mark things which are highly confidential and have specific data policies that actually protect those. So that's how you secure AI. And finally, again, as this digital workforce starts getting built out, you can imagine you can start with like tens of agents, hundreds and then get into thousands of agents across your teams, platforms, applications, et cetera. And that becomes a nightmare for the CIO, for the IT leader to go manage.

And this is really where for several years, IT leaders have relied on the product called MuleSoft to manage their APIs, which connect their business. And that's what we are evolving into this product called Agent Fabric. It's, think of it as a single registry that helps you manage all your agents, not the ones that have been built on Agentforce, but they're built on any provider. We basically are able to scan; monitor them and do things like FinOps to really look at what value you're getting of the agents versus what you're spending on them. All right. Let me actually jump to the demo and actually show you specifically Agent Fabric and Salesforce Guardian because there's 2 very important, it's not very obvious. The sort of the boring tasks that need to be done in enterprise, but it's extremely important that you get this right.

So this is basically the homepage of Agent Fabric. And as you can see, we essentially tells you there are 726 agents that are running. And what you see are all the scanners. So these are basically tools that are running continuously in your enterprise to discover new agents as they're coming up, right? So, and if I go take a look at the Agents tab, let's go to the list view. You see on the provider side, agents from Azure, AWS, Google, Agentforce, so all the providers. And if I want to go deeper into any specific agent, let me just pick 1 of them. It's the recruiting agent. This was shown in the demo yesterday, right? As an admin, I get a bird's eye view of all the skills that the agent has. Like this is important because you really, it helps you understand the logic that the agent is using to come up with decisions.

As an admin, I can go deeper, look at lineage. Here's where you see the integration across products. So this is the grounding data, the enterprise context that this agent is using. And all this information has come from Informatica. Remember, Informatica helps you discover all your data assets, and that has been integrated into Agent Fabric. If I'm an admin, I can actually map these data sources onto the agent, and I can look and say, hey, is it hallucinating? I can come and debug things over here. The other important thing is you want to monitor these agents as an admin like what's the average latency? If these agents are front ending your customer experience, then you really want to know how long does it take for the agent to respond, what's the error rate, how many policies have been violated.

And then what's the total number of requests? Like how much of volume is being sent to the agent because that will be directly related to your cost, right? So here, as I see the agents, the specific agent is seeing a spike in the number of requests. So let me go dig into the cost management piece of it. And here's where Agent Fabric does a really good job of enabling the IT leaders to manage their budget really well. So you can think about creating sections of your budget, if you will, for a certain class of agents, and that's the maximum amount that they're about to spend monthly, and then you can manage things like that. Another interesting thing that Agent Fabric does is that behind the scenes, it actually observes the work that the agent is doing to come up with recommendations in terms of like what changes can be made to make this agent a lot more optimized in terms of token usage, right?

So that's directly impacting the cost. So in this case, I can go to the agent I was looking at. There's all these optimization opportunities. And obviously, you have to be careful that these don't change behavior of the agent. But assuming that's true, that it doesn't, then I just go apply and that's it. Basically, that's very simple. So you can manage and govern the agents, you can look at their performance and the value creation that's happening and manage their cost. It's going to be very, very important as the volume of agents go. So that's the IT leaders' part of it. Now let's take a look at Salesforce Guardian. You'll see like the way we are trying to create this is you have this notion of a security score, which is overall, like what does the security health look like, how many assets that you have across your enterprise contact center I mean monitored here, you see it's about 94% coverage.

But the 2 very important things when it comes to agents are essentially agent identity. So as the agents do their work, there we actually observe every action that they take. I imagine the scale at which these agents are running. And based on the actions the agents are taking, you determine whether the agent carries risk or not. For example there, the agent was supposed to do X and it's doing A, B, C. Well, that's, they're deviating from what they're supposed to do. So we capture those observability logs and then analyze that and determine the risk of the agent. So you see in the dashboard over here of all the monitored agents with different levels of risk, 2 of them have been called out as high risk. That's something for the security to take action on.

The other important part is the data assets, which is your enterprise context was built out of a certain volume of data, how much of that data has been classified and understood, right? Because if that percentage is low, then you land, there's a lot of risk that these agents might be touching data that you don't want them to touch. So anyway, that's just a bird's eye view of how Guardian works. So Agent Fabric for the IT leaders, Guardian for the security leaders. So beyond just selling you the agents, it's this whole manageability piece, the security piece that lands up becoming very important. All right. So we can go back to the slides. Yes. So all the stuff that I just showed you is things that we have today, right, which all these products, there's a method to the madness. It's not just a random suite of products that we have.

The journey that I walked you through is exactly what an enterprise would do. Now the interesting thing now is I'm going to pivot a little bit into the future of this Data 360 layer that you saw in the stack, how that evolves into what we're going to call the enterprise AI harness and what's our perspective over there. It starts with the frame essentially is every role in the enterprise is evolving. I mean we've sort of spoken from the perspective of the business users. But like I said, the IT leaders are going from just managing devices and apps to sort of managing these agents. You take a look at the data leaders, they're not just managing data warehouses and lakes, creation of that enterprise context that makes the agents accurate and efficient is going to be a very important part of their job.

This is things like that, that most of these roles, things are going to change. So what is an enterprise harness? This term has sort of become like 1 of these platform terms, right, which is extremely confusing. But fundamentally, this is the link between what your AI models, the intelligence that you have and then the understanding of your business. The harness is essentially what brings it together for the business outcomes that you desire. Now in Salesforce, like this is a stat that we showed, the bottom layer Data 360 essentially is going to evolve into this Salesforce's version of the enterprise AI harness. And it has 6 capabilities. I won't sort of go into the depth of each of these and an AI control plane that brings it all together. But fundamentally, today, there is a lot of discussions around harness to protect a single agent, right?

So where does the agent run, the run time of that, the evals, which basically tell you the agent is efficient or not, right? You test; you make changes to the agent. And then depending on what responses it's giving, you manage that against your evals, that's, think of that as your test set. And that's, those are the only 2 things that enterprises are focused on. We believe there's a much larger play over here, right? Starting with trusted models, which is we wanted to give you a choice in the models that you pick, right? It's not, you don't need the most extensible model for everything that you do. As a platform, we need to figure out based on the outcome that the customer is seeking based on price performance, what's the best model to pick. Trusted context is everything that I spoke about, which is synthesizing your business understanding, your semantic models, all of that sort of stuff.

Agency and actions is how you, when a customer describes the outcome that they are seeking, the agent and the model they create a plan. The sequence of steps that need to be taken to execute that decision-making comes in the agency framework. And then the actions, the trusted actions is how you execute each of those steps. And every step that gets executed has to be governed just to make sure that the agent is not doing anything that it's not supposed to do. And then trusted governance is essentially getting all your data policies right because that's going to influence how your trusted context gets created and security is about agent identity and data protection, right? It's a very comprehensive view that at Salesforce, we are taking in terms of harness definition, not just protecting the agent, but if you can wrap your arms around the harness, you should be able to bring that with the headless APIs to the AI of your choice.

This is really where we are sort of going towards the future. And an example of the model choice is what we announced in the keynote yesterday. Koa, it's actually very exciting. It's the very first CRM reasoning model from Salesforce. We basically post-trained the open-weight Nemotron from NVIDIA. And like 27 years of deep product understanding and appreciation of the CRM workflow, more than 10 years of data and AI research, it's sort of been baked in. So there's a lot of value, very excited. And this model has been designed for long-running agents. And what I mean by that is agents that can actually create a chain of thought which has multiple steps and make decisions at each step. So you can imagine a complex sales opportunity where what's the next best step to take to go make that happen or a very complicated service case that an agent is dealing with to get the customer to a better state.

So complex situations like that, Koa can greatly help. Very, very excited about it. And then finally, evolving Agent Fabric into this AI control plane, which essentially helps you discover all your agents, govern them, connect multiple agents through standard protocols and eventually do your FinOps, the cost control, all of sort of coming together. That's a big part. Now 1 important thing is it's a very open and composable system, while Salesforce is going to have a very comprehensive answer for each of these challenges because we believe without having a complete solution like this, production deployment of agents is just not going to be successful, right? That's why we are taking a very holistic approach here, but it's open and composed. We realize that like our customers may have a different security vendor, a different data vendor or a different vendor for the AI control plane, all that's fine because through our headless API, we are able to integrate with each of these partners.

So that's a big design point from the very beginning, saying we have a complete solution, but you don't have to use everything from Salesforce. And the thing that I'll finally say is while this is, we are sort of evolving this into the harness, the approach we are taking in terms of building this product is very incremental. So if you are a Tableau customer, if you're a Data 360 customer, if you're a core platform customer of any of our clouds, Sales, Service, et cetera, MuleSoft, you already have taken steps towards actually getting into this harness because we're breaking down the capabilities of all these products and bringing it together in a very composable and a unified way. It's pretty exciting. I think this is something which is going to be very important for our customers to do their Agentic transformation right. And would love to hear your thoughts on this after the session. So thank you.

Thank you, Patrick and Rohan. The demos were sensational. Next up, we're going to have Miguel and Alexa. And I want to just jog your memory. 1 year ago, Miguel was on stage. Miguel was our Chief Revenue Officer. His remit has expanded. He has all of go-to-market motion now. That includes customer success and partnerships. And along with that, Alexa's role has been elevated. And how about a big round of applause for Alexa now, our new President and Chief Revenue Officer. And I couldn't think of a more critical time to have you on stage. So thank you.

Thank you, everyone. Rohan, Patrick, thank you for teeing up so nicely for us. Alexa and I are the lucky executives that get to take this amazing software stack to market. I was here a year ago, same exact location. I think it was October in front of many of you. And Robin and I, we were telling you how excited we were with the momentum that we were seeing in the business, the massive opportunity ahead of us with the Agentic Enterprise and how we were investing heavily and how we were investing heavily and wisely to capture the opportunity. The interesting thing is months later, we started hearing this chatter, this noise in the market, something called SaaSpocalypse or the false narrative that AI was killing SaaS. You know what, we are very competitive. We are like what the hell is going on. This is not what is really happening.

I mean, luckily, the market are starting to come to its senses. And fortunately, I think by now, all of us, I think we believe that it's time to put this false narrative to bed because the reality of the business that we've been operating on for more than 4 quarters already, more than a year, is like the piloting phase of AI from 2 or 3 years ago came to an end. That's what we told you and people were going in production at scale, betting on the big software platforms like us. That's been the case throughout. So we call this new phase SaaSceleration. Some people call it Renaissance, my accent is not great, but you get my point, right? The message here is AI is not killing SaaS. In fact, AI is making software vendors like Salesforce more relevant than ever because AI needs the trusted context and the secure governed deterministic execution to drive value in the enterprise.

I mean that is fundamental. And we have all the pieces of that software infrastructure that you saw earlier. I'll give you a small anecdote. It just happened to me in the last meeting before coming here, I was meeting with 1 of the largest PE firms in the world. Okay? There aren't that many in this category. And 1 of the leading partners, he was telling me that 300-plus large portfolio companies. 75% of them are on Salesforce. And he told me something that I wrote down a piece of paper that was pretty amazing. He said, Miguel, I just want you to know that the Salesforce platform, your business application is the most prolific, prolific of all the platforms. And he named SAP, he named ServiceNow, he named Oracle, he named Workday in terms of driving value with AI to our portfolio companies. And I'm like, you're going to come on stage with me later and say that, but he's conflicted because he's probably 1 of you, he may be here, I don't know.

But the net-net is it's a really renaissance era for us. These are the 3 key messages that Alexa and I would like to share with you today. Number 1 is we see increasingly strong demand in this area. AI is accelerating SaaS, okay? And we are, we continue to invest aggressively, again, wisely also in the top AI markets to capture this opportunity, okay? Very similar message that Robin and I delivered last year. Second, this is a bit new because we were learning last year. We, and Alexa is going to go in detail on this point, the second point. We have now many different levers to monetize AI. We identify a bunch of drivers that are making our customers buy packaged solution with AI embedded. And we also have multiple commercial frameworks to monetize that opportunity. We are meeting customers where they are.

Not every customer is in the AI transformation journey at the same place, but we have a pricing, a commercial model for each of them. And then finally, hopefully, I mean, I was very, I was worried that you guys came today for this event and then left and you were not at Monday, Tuesday, the keynote. I'm very glad that 2/3 of you were at the keynote. Hopefully, you've talked to a lot of customers. I talked to a lot of customers myself. At every meeting, I have to pinch my cheek. I'm like, 'Oh my god, this is so incredible what the customers are telling me.' The proof is in the pudding, and you've heard from many customers on the keynote, et cetera, you're going to hear from 2 more customers. We heard from Frederik, we're going to hear from the Adecco Group. We've heard the marketing story. We're going to hear the real story here by their technical team.

And then also, we're going to hear from Anthropic, and it's super exciting when you hear from customers. Now on the first point, I'll cover the first point. I know that you like this slide on the left side from last year. So Robin and I, we were very confident about the business. We knew that things were going very well, okay? We had really strong and healthy succession of quarters, okay? And we told you last year that we were going to see an organic subscription support revenue reacceleration within 12 and 18 months, okay? Guess who said 12 and who said 18, okay? This is like a partnership. Well, it's been 12, 13 months, okay? So we have guided Q3 subscription support revenue. If your models work correctly, you'll see that there is the beginning of a reacceleration. Why? Because of what we told you last year is the chart on the left side, the magic word, net new AOV growth.

Net AOV growth at that point when we were here, the lines were crossing. We knew they were crossing, okay? They did cross in Q3. So the net new AOV growth outpaced the AOV growth. Mathematically, when that happens, the AOV accelerates. AOV always grows, obviously, but it was decelerating for years. Why? Because there was a lot of, because of this. Because we have some years after COVID where the net new AOV was actually even negative growth, and it was pulling down the AOV growth. Now since Q3, the last 12 months, the last year, net new AOV growth has outpaced the AOV growth. Okay? And we feel very confident in this net new momentum going forward. Now look at the business metrics. I mean, most of them, I think all of them we've already shared at our earnings.

cRPO, not a bad number, 14% we delivered in Q2. We guided 14% this quarter for Q3. That 14% for Q3 does not include the Contentful and Fin acquisition that we just closed last week -- this week. Pipeline at record levels. That's a little bit of a new info for you guys, but it's in the very high teens, very healthy coverage ratios. Coverage ratios are healthier than other years with increased commits, obviously. It's looking good across the board. Okay, across segments. Contract lengths are increasing. I mean we gave you this stat in Q2. We do 100,000 transactions per quarter, give or take. A lot of them are renewals. Many of them are new bookings. It doesn't matter whether we're talking renewals or new bookings. We give the average of everything. It doesn't matter the segments. We took every deal band, contracts were expanding in every deal band. That's good. That's good.

Then you add the fact that seats, sales, seats and service seats. Remember, the world in SaaSpocalypse world, the world is coming to an end. There will not be any more sellers. There will be not more service agents. Guys, you have a seat-based model. Salesforce is going to cloud. We haven't seen still any decline on number of seats for sales or number of seats for services. And we track that obviously very closely because guess what, at some point, there might be. But we are not even worried if when that point arrives, which so far hasn't arrived, it's still growing. We have so many new commercial frameworks to monetize the AI opportunity that, we'll be okay. We'll be okay. But so far, it's increasing. And of course, Slack seats are flying through the roof, okay? It's the hottest product right now, I think, in enterprise software, definitely within Salesforce, but I think in enterprise software.

And then the consumption, AWUs are exponentially growing pretty much every quarter, we do as many AWUs, Agent Work Units, which is the closest thing, it's not tokens, it's the closest thing to productive work delivered to the enterprise. We do more in 1 quarter than in the history until now. I mean obviously, mathematically, that will have to stop at some point. But it's a lot of consumption, a lot of customers, Agentforce customers consistently consuming. So, and then, by the way, I didn't mention something that I think it's going to be 1 in your slides, Robin, every single AI company, the same AI companies that were going to kill SaaS and replace all the CRM systems, all of them are going fully into Salesforce. In just 1 year, in this AI segment, we multiplied by, actually, it's 100%, nearly 6x fold the business that we do with them. Not bad.

You take the top 9 AI companies in the world, the top 10, 9 are wall-to-wall Salesforce, okay? The 10th one, probably you can imagine who the crazy company is, it's an amazing company, by the way. They are actually not wall-to-wall, but all top 10 run on Slack and 9 out of 10 wall-to-wall Salesforce, okay? And they're multiplying by 6 the investment that they're doing in us on a year-on-year. These signals, these business signals are clearly not a SaaSpocalypse situation. So I'm not going to use the word SaaSpocalypse anymore. I hope you don't either. And the market will come to senses this little by little, and we are ready. Now because we see this momentum, because we are excited because we are talking to customers, because we're seeing the demos because we're seeing the new surfaces because we, I, Alexa, all our teams, we are now working.

I mean I go to work, I think I'm going to a video game. I have my, I mean, if you guys see my surface where I work, you don't believe that this is a serious company. This is -- I'm having so much fun. I have a screen that looks like Patrick's demo. By the way, in your, when we finish this session in the break, 1 person in my team is going to show you my cockpit, which is a skin that I put on top of Claudeforce, okay? Because the good thing about Claudeforce is you can say a few things to Claudeforce and it builds a new skin on top of it. Okay? And until yesterday, I couldn't demo, definitely not to the finance community because particularly this is a public event, it is broadcasted. And you will see all my numbers, the numbers for Q3, which I'm sure you're super interested to know. And all of them are going to be on my laptop.

Ning, who works in my team is going to show it to you, but we told something to Claude this morning. Create a toggle button on the top right that says demo mode. So when she clicks that, everything that is data in our metadata model, everything that is data gets blurred. But she's going to do from time to time like this so that you see that this is real data. So this is not, because I think, Alexa, your demo is going to be dummy data, okay? It's on the Slack, there's another surface. But it's the same system with dummy data because you didn't ask Slackbot to do that. But if you do it, so, but listen, it's fun. And if it is fun for us, it's fun for our customers. I mean, Rik, I mean, he's just playing with it. I mean I feel bad for him because he's just been playing with it for 2 days, and he was here in front of the world doing a demo.

And this is very basic, what he just built. Imagine if you build that, you iterate 1 week and you have the most amazing dashboard in the world, driving value to you, sending you signals. Every time that you put a prompt, you are creating an agent that does things for you. So I don't log into Salesforce anymore. Parker, thank you very much for leading the way. I think it was here when you said I don't understand why people will log into Salesforce in the future. You were totally right. I was sitting there and I'm like, this guy is crazy, okay? He's my age basically, but white hair, I'm like this guy is becoming crazy. Why he's a visionary? And I pretty much 1 month later, I stopped logging on to Salesforce. And I use Salesforce 100x more, okay?

Because of all this excitement, we are tripling down, investing in capacity, also under my new role, investing across the success area at these builders, I have 20,000 more people in our revenue because we're bringing the whole go-to-market together. And I gave all of them 1 focus. And you guys can imagine which focus I gave, which objective I gave to those new 20,000 people in customer success, professional services, partnerships, 1 metric. You can speak. Not you. Look at the chart, net new AOV. And why net new AOV and not customer success? Because net new AOV means it's the ultimate metric of customer success. If customers buy more and are happier and adopt more, which is net new AOV, okay, it's good for everyone. So anyways, things are happening. We are doubling the capacity. We've added 30% more AEs in the last 2 years.

And at the end of this year, we're going to have double-digit growth, 10%, 12% growth, more AEs in the core countries than 1 year before. We're also adding 1,000 builders as we speak. So we are very confident in the momentum and the direction and the net new momentum going forward. And then we have, we are very confident about the software infrastructure. Now one; remember the words from the PE person. This PE person represents 180 companies. And these are not Mickey Mouse companies. They are pretty big companies because this is the biggest PE firm in the world, 1 of the biggest. We are the platform that drives more value to the end customers through AI because people want to consume AI through packaged applications. This is the trend. We take care of all the complexities.

We take care like Claudeforce. Everybody can connect MCP servers and build their own MCP servers, but you want to do it securely. You want to do it with 0 data retention. You want to do it with all the skills that matter. So that's the power of our platform. We are very confident. And with that, I'm now going to hand over to Alexa. We already applauded her. I'm so proud to work with you. She's amazing. She's a force of nature. And it's my time, my time.

I didn't get walkup music. I was very; I was sitting in that seat wondering what it might be. Thank you, everybody, for having me here today, and thank you, Miguel, for the introduction. You guys don't know me yet, so maybe I'll share a quick personal tidbit. I grew up in a family where we had a motto that was per aspera ad astra. So through difficulty to the stars. And I think it was really appropriate for what we've experienced over the course of the past couple of quarters because difficulty, I think, sometimes is a necessary condition for success. And I think what it did for us, as painful as it might have been, was forced us to get really clear on what our customers needed from us in that and also what we do incredibly well.

And I think at the intersection of those 2 points is we discovered that our customers had a huge amount of trust in us and our customers needed us more than ever to help them transform. So that's what we've been really focused on. We're sort of ignoring the noise and focusing on how do we go execute to drive against that customer expectation of transformation and trust. And so candidly, I'm even more excited to step into this new role in a period of, we won't use the old word anymore, but in the SaaSceleration or the Renaissance. But I think that difficulty has been a huge learning aid for us. So let me sort of help you understand a couple of things. 3 things on this chart, although it's going to be a pretty dense discussion of those 3 things that I want you to take away.

The first is that we think we have an incredible opportunity to go monetize AI. The TAM is exploding. I think that the word we will not reference anymore did not incinerate opportunity for software, it expanded it. And so Robin will talk a little bit in her section about some of the hard numbers and how we see that spread out across industries and customers. I think 1 interesting way to think about it is the expansion of knowledge workers. There are, give or take, roughly a billion knowledge workers today. We have 50 million of those knowledge workers roughly on our platform. There is huge upside to go capture a really significant portion of the market. And that is before we ever think about an agent as a knowledge worker, which I think is going to geometrically expand that knowledge worker base.

There is so much that we can go do here. So that's the first piece. I think the second piece is to think about how are we going to put our capacity. So Miguel talked about all of this capacity that we're investing into the market, double-digit growth from the builders. And you're going to see us, I think, do a couple of things. The first is recenter on the markets where we think AI is going to grow the fastest. And so I think top markets has always been pretty central to our strategy. I've been here for a decade, but we're going to double down on the 9 markets that we think are the biggest and are going to grow the fastest where we have infrastructure to go in and accelerate. It doesn't mean we're not going to invest in the rest of the world. It doesn't mean we're not going to look for opportunities to move really quickly and find upside potential that we haven't tackled yet, but you're going to see us concentrate our, put all the weight behind the arrow and go concentrate our effort and our energy.

We're also going to be really thoughtful about what does the capacity mix need to look like. You saw the stat that we've scaled we'll exit this year scaling to 1,000 builders. And so as I sit with customers, and I spent, I probably met with, gosh, I don't know, 150, maybe getting close to 200 customers this week at Dreamforce, 75 CROs yesterday. The #1 theme coming out of -- I spent 2 hours with some of the top CROs in the world. The #1 theme coming out of that session was help me build this. So they saw Claudeforce, the demo that you saw Patrick show earlier. They also saw a version of how we run our business in Slack. There will be other examples of that on other surfaces they want and need us to help them build that.

And so part of the capacity modeling that we're doing right now is thinking about, I don't want to show up to a customer or anyone on my team to show up to a customer with a PowerPoint. I want to show up and build a prototype; I want to push that into production as quickly as humanly possible. So that's sort of a thought in terms of how are we mapping our capacity to the demand function that we see. And then I think in the spirit of how we started, we're being pretty thoughtful about, okay, what are the demand drivers? What are the patterns that we see. And what I would tell you is this is indicative, right? And you guys are talking to customers, and I think you recognize that there's alternate patterns, but I think this represents probably the most common one. So let me sort of walk you through what I see more often than not as the way customers adopt the platform and how we're sort of thinking about driving that monetization motion.

So the first, when I talked about knowledge workers, is we unlock new categories of knowledge workers. We unlock additional seats. Miguel talked about, we are really confident because I see seat expansion without even uttering the word AI, I see seat expansion as an opportunity. And I think a good example of that is Replit. Everyone in the room is familiar with Replit, 1 of the fastest-growing digital native or AI native companies. They scaled from $10 million in ARR to $0.5 billion in ARR. We scaled with them. Those are core sales seats that we landed into the account, and we're growing as they grow. The thing I think that's really exciting is it's not just sales, it's not just service, it's also Slack. You heard the story about Adecco on the main stage. We're going to tell you the real story today, so I'll skip that one for now.

But I think the net message I will leave you with, and I think I heard this in the CRO Summit yesterday, too, people still need to build contact centers. They still need to build out pipeline management systems. They're still going to build marketing campaigns. So there is a big opportunity here for the core portfolio. I think though, and Replit is a great example of this, as they expanded their core sales and service seats, they started to think about agents. And so the most recent opportunity we had to engage with them, they expanded into agents, and the place that I typically see clients start is in this employee agent use case. And so they want to build out capabilities to make just like we think about productivity, their own employees way more productive. I was with Bouygues, one of the largest European telcos in Paris, probably back in March, and I did a ride along.

This is really fun. I grew up in Canada. I have a grade school Canadian French. They did the entire meeting in French, which severely tested my capabilities. But we did a ride along in their contact center. They have 6,000 people sitting in a contact center just outside of Paris. And they have implemented Agentforce for internal use cases for those 6,000 contact center employees to help make them more productive. They wanted to stop swivel-chairing, and they wanted to be able to help automate the process of sending text and e-mails, searching knowledge articles, first and foremost, 600 knowledge articles, but also doing outreach to customers who are inbounding the contact center. So I think it's a really powerful example of how customers can start with their internal use cases, but they don't stop there.

I think once they understand the power of an agentic capability, they start to think about, gosh, I have to get that in front of my customer. And CrowdStrike is another really good example. I've worked with CrowdStrike for 5, 6 years. They started their agentic journey as a Sales and Service Cloud customer. We implemented a whole bunch of agents, but the 2 biggest agents were an ask legal agent and they follow MEDDPICC, so they do sort of like a qualification process, and they wanted to help boost their sort of pipeline quality. So we launched those 2 internal agents for them. Fast forward, they now have multiple agents in production, but they now run an external agent, their service agent, and that is containing 83% of the inbound requests that come to their contact center. You can't actually file a support ticket without touching Agentforce when you talk to CrowdStrike.

So I think that's a really powerful example of the pattern. And as customers do this, and you heard Rohan talk about it, invariably, what they recognize is they need to harness their data to make sure that they're high grading those agents effectively so that they can serve their clients and they can serve their employees. And so FedEx, another client that I spend a ton of time with, they produce an astounding amount, a petabyte of operational data a day. They have 600 different, actually more than 600 different data sources, and they're using Data 360 to kind of harness all that data and bring it together. So their 2 largest data sources or the data lake environments are Azure and Databricks, but we bring all that data across those 600 sources into internal agents that are helping them do a bunch of things.

But 2 of the most powerful things they do is look at dormant accounts and look at quotes that have been abandoned. That process took them 3 or 4 hours, but the ROI has been like tremendous off the chart. So I think that's a great example of kind of how the discipline around agents ultimately drives the motion towards we need better data. And as they build out more and more agents, they got to think about governance. So ADT is another customer that I spend a lot of time with. I'm giving you a lot of bright and shiny ones. So let me give you 1 that was a little difficult. Maybe 1 year ago, July 4, actually, I got a call from Marc saying that he'd heard from Fawad, who's the COO at ADT and that he wasn't happy with the Salesforce implementation. And so we obviously parachuted the team in.

I think Mark Sullivan got the same phone call. He's had a history with Fawad at State Farm. We had to do a bunch of work to rebuild their data model because what they wanted was to be able to go way faster. They were swivel chairing across a bunch of applications. They wanted to implement agentic technologies and the apps or the orgs that they built just weren't ready to go do that. So we suitcased our team in, great news. They've sunset a bunch of these legacy applications. Everyone is standardized now on Service Cloud. So again, an example of number 1. But they're starting to think about or they're working with us deeply to partner on, okay, as I build out all these agents, how do I do 2 things really that are really important to them? How do I govern them?

How do I understand where my agents are? How do I register them and kind of manage their behavior? And how do I manage their costs? They're super focused on managing their OpEx responsibly. And so I think this is a huge unlock for them as they continue their agentic journey. And then finally, and this is a little bit of a departure. It is a super new use case. I'm being sort of transparent and probably leaning in a little bit to an opportunity that we see from a monetization perspective that's really, really new. I love it for a lot of reasons. A, we didn't see this kind of business opportunity at, call it, the beginning of the second quarter. So in the last 90 days, this has shown up. And what we're seeing now is Slack emerge as the agentic operating system for agents.

And as people launch their agents into Slack, what they need is a super high latency performance pipeline to make sure that they are fueling their agents with Slack data. And so this is something that I am really excited about. We have probably, the 3 deals that we've closed, a bunch more in the pipeline. Think about digital natives as really being the target area for this. And the thing I love is that it breaks the association. Miguel talked about seats. I feel really comfortable about the seat expansion opportunity that we have, but this breaks that association. I monitor every digital native and how they use our Slack system today. And I think 1 of the really exciting things as you look at it is that the correlation between API calls and usage doesn't map to headcount.

You can have a really small but fast-growing company that is using, that is building agents like crazy, and they don't need to have a lot of employees. So I think that's really exciting. And then finally, we think about more ways to buy. And so you might say, no, actually, you guys already have a lot of ways to buy. But I think this is really important for us, right? We need to meet our customers where they're at and make sure that they've got very flexible, very low friction ways to onboard onto the platform. If you look at Q2, 7 of our 10 largest deals were AELAs. And I think as we start to launch Salesforce Commit more fully across the team, that's going to be a major driver for revenue acceleration for us as well.

On a final note, and Patrick, I think, showed you this slide, the premium upgrade motion, I think, is going to be a huge acceleration vector for us. So for those who aren't familiar, I'll just show you, you saw this before, so let me orient. We haven't introduced a new value edition SKU in over a decade. And so in the last, call it, 8 quarters, the team has been focused on how do we go take that base of core CRM users and move them up the value chain. For every 1% of the base we move, it's $100 million. But more importantly, I think it's an opportunity to push all of our customers to fully leverage the rich capabilities that we're offering in Max Edition, including headless. So this is something that every single seller, every single person in the company has been enabled on. They understand this is the call to action. I have not walked into a room at Dreamforce and not asked people to upgrade.

So if you guys are excited about upgrading your companies, you can scan the QR code and learn more about Max. But I think this is going to be really powerful, and we've oriented our partners and our company around driving that upgrade cycle and making sure that we work with customers to make them successful on it. I think just a final note, and you heard me talk about the 8 quarters. In the 8 quarters would, frankly, I would tell you like partial emphasis on this, we've built out $1 billion of AOV. And when you look at the sort of mechanics of it, in the upgrade pattern, you see about a 60% to 80% uplift depending on the customer and which edition they're coming from.

But when you look at the individual sort of account level data, what we see is a 1.4x expansion of the ARR. I think this is really powerful. A, we're growing, and that's kind of the key message. But when customers renew, you guys can do the math and kind of math, the ARR is slightly less than the upgrade fees. They're remixing the portfolio. They're realigning to where they need to go and kind of fuel their growth or their transformation efforts, or the products that they need to fuel their growth and transformation efforts with. And so not only does it lead to top line revenue expansion for us, but I think it also mitigates any fear of attrition in the future.

So I think this is really powerful, and we are going to put the whole force of the company from marketing and campaigns through enablement, plays, programs, incentives and calls to actions in the field. This is the execution point. Everyone is going to be focused on monetizing the value edition. So I'll end there on the monetization note. I think there's been a lot of discussion around Claudeforce at this conference. And I think 1 of the things I'm hungry for is to show you a little bit more about how we're using our own tools, Slack included, powered by Anthropic, but Slack to drive our business.

Claire is going to help me here. I was wondering if you're in the back actually building a demo toggle into the Slackforce demo. But before, while she's setting up the demo, a couple of things. So we're obviously very focused on driving growth, increasing our sellers, improving their productivity. 1 of the ways that we do that is through sort of revamping from the ground up our sales processes. Forecasting is 1 of the processes that we know people spend a ton of time on. So we have totally rebuilt that function across the go-to-market organization since the beginning of Q3.

We rolled out, and let me know, Claire, when you're all set. I'll do the Patrick riff, see how long I can do this for. We rolled it out at the beginning of Q3. And this is what we use to run the business now. You're not going to see live data, as you heard Miguel talk about, you're going to see dummy data. But otherwise, this is exactly what we use every Friday when I gather my leaders together to run the business. So I think it's really powerful. How many people have seen Slack before? How many are familiar with it? Okay. All right. A few more things to me. You guys weren't going to have seen Slack, you guys, you know what you're about to look at. Are we ready to go? Okay. Do you guys mind switching over to the demo? Okay. Great.

For those of you who haven't seen it, this is a surface in Slack. I think this is 1 of the most powerful -- I use Slack to do 9 million things. I use Slack to prepare for today. I use Slack to prepare for customer meetings. I use Slack to collaborate with my team. I use Slack to run my global forecast process. I'm going to point out a few things here that I think are really important.

So the first thing that you're looking at is a dashboard. Everyone in this room is familiar, but there's way more to it than a dashboard. I look at this every morning when I start my day, and again, this is not real data. I look at this every morning when I start today, my day, to kind of understand, all right, what's happening in the business? What's changed overnight? What should I spend my time on.

There is no more precious commodity than time. I want to know where to orient it. And so this does a couple of things. It tells me where we are at any given point in time, and it tells me who are the drivers behind it? Who are the people who most materially represent the opportunity that we have in any given quarter. You can see that here.

It also calls out, look, where might you have some exposure in the business? Where do you need to go pay attention? This is the what to watch. Miguel talked about the fact that our pipeline was up this year. And so I think that -- if I looked at this and saw that my coverage was high, I think that's basically driven by momentum, but I'm still going to go double check that to make sure that we're operating with the same quality of pipe that we typically do.

And so this kind of gives me a good orientation for what's going on, what do I need to go focus on. And then I might dive more deeply into the forecast by leader. Well, actually, I'll show you. So this is just, again, dummy data, so it doesn't necessarily show you a pattern. But I would look at this to say, okay, what's happening? Is there a trend that I need to understand here from a macro perspective? Is Europe doing really well and EMEA is slowing down and Asia Pac has a problem.

So I can look at this kind of global view. But I tend to spend most of my time in the leader view. And so I might pick as an example, Lenore Lang. Lenore Lang succeeded me in my previous role, so she runs tech media, telco and our data foundations business, our largest operating unit by revenue and people. I might pick this business and say, okay, tell me about Lenore's business and what's happening.

And so I can see here that Lenore is calling $172 million. Slackbot thinks that's high. Now in this case, again, not real data, I'd be like, look, I think Lenore is going to crush it this quarter. So -- but it will give me a view of what Slackbot thinks and why it thinks this is going to happen.

If you sort of -- this is obviously instructive and helpful. But if you think about it, I have teams of people who spend a lot of time. Forecasting is really important. You depend on me to give you an accurate view of what we're going to do. My team depends on that, too. And so we spend a lot of time making sure that we deeply understand the numbers, but I think Slack can help us do that better and faster.

And so I can actually dive into for each of my leaders and I'm going to stick with Lenore here [indiscernible]. So I'll tell you this is a real scenario. It's not -- it didn't happen with Lenore. It happened with somebody else, and it wasn't this deal that you see here, but it was a real deal that happened last week on the forecast call.

And so I can go through and I can understand her right. This is Lenore's -- like the overview of Lenore's business, here are her top 3 deals. And I might see that she's got a big deal at a tech company that is at risk. And so we'll talk about that on the call, and she may give me some feedback about why that deal is struggling.

And one of the things that I want the team to do is get Marc involved right now. There isn't a deal in tech that shouldn't have Marc connected to the CEO. And they are not eager to do that. They want to be perfect. They want everything buttoned down. I want them to do it fast. I want Marc to reach out to the CEO, so I know that if something is wrong -- when he executes that reach out, I'm either going to get good news or bad news, but I'm going to get it fast and I prefer it fast.

And so I'm on this team's case right now to go and get that outreach. But I can also look at, hey, what's happening in this deal right now? Are they doing SICs at Dreamforce? What are the exact connects that they have happening? And then I can actually go into it and kind of say, "Hey, team, let's get this down or let's get this text message set up for Marc and let's have them reach out to the account really quickly."

So that's just kind of a very quick run-through of how I'm using Slack powered by Claude, powered by Anthropic, excuse me, to run the business. And a final note, this is Miguel's favorite example, so I promised him I would do it. I will be in London later this month. And one of the things as we sort of make these trips into the market that we want to make sure we're doing is spend time with customers.

So I can ask Slack, what are the top deals in London in Q3. And it will go through and it will search. And you'll notice it's going to come back. I didn't tell it how many, but it's going to come back to me with, I think, 7 deals when I checked it this morning. That's going to say, hey, here are the top 7 deals in London.

And so further, I can say, look, I want you to send a Slack to every one of the opportunity owners for each of these big deals, copy their entire management team, tell them I'm going to be in town and ask them what I can do to help get this closed in Q3.

And so you'll see just down at the bottom here, it comes back and says, okay, do you want me to send it as a DM? Do you want me to post it in the channel? Do you want me to curate it so that it's fit for purpose for that individual team? I say, yes, that's what I want you to do.

So super quick example, but I think it's indicative of the transformation that we're driving as Customer Zero of this technology. And I think the opportunity that we have to go help all of our customers transform. So with that, I'm going to pass it back to Miguel, who's going to share 2 real-life examples of customers we're transforming right now.

From the Salesforce executives. And now we're going to hear from the real heroes that are getting this incredible stack and driving enterprise value in their companies. So I'm going to start with Adecco, okay? And by the way, this has been a little bit impromptu because we saw Adecco many times. We talked about Adecco. The CEO was with us at the keynote, great marketing messages.

And then right after the keynote, I met with the working team, and they went through all the progress. I typically meet with them every -- about 3, 4 months. And I'm like, wow, this is even better than I anticipated. And I had a good conversation with my friend for many years, Pierre Matuchet in the IT area, Senior Vice President of IT now Pierre join me.

I think they're going to put a chart for us. Give a round of applause to Pierre Matuchet, Adecco Group. There. And honestly, the reason he was -- his eyes were like on fire when he was telling me the progress that he was driving with Salesforce and I'm like, you know what, what are you doing tomorrow night -- tomorrow afternoon, I'll say, tomorrow night also I invited him today to the concert.

But what are you doing tomorrow afternoon because I'm meeting a few of my friends, and I would like you to be with me and tell your story. It's funny because Denis, the CEO, arrived at the meeting at the end, and I said, Denis, I need to get your approval to get Pierre to be with me at this meeting, and he said, yes, so here we are. So thank you so much. Let's take a seat.

So Adecco, I don't know if you -- hopefully, you know Adecco, but it's one of the biggest, if not the biggest staffing global company in the world in HR services, about 34,000 people in 62 countries, about $25 billion of revenue. And we are so grateful that you're here, to your CEO, to the partnership that we've had over many years. And the first question, if you remember last year, I always ask the question similarly, what was our relationship between Agentforce? Sorry, before Agentforce.

Before Agentforce.

Yes.

We were using your CRM for 15, 10 years. It was good. Nothing spectacular. We have a relationship.

Okay, how do you like the [ interview ].

Not a strategic relationship, not a strategic partnership. We were a bit stuck from our side with 42 instances of Salesforce, so not easy to manage, maintenance cost and so on. We need to build some complex mechanisms to have a view of everything. And we get more and more requests from global customers like Amazon, like Siemens to have the same process everywhere in the world to be able to give them a global reporting.

And we were in Dreamforce and working with your team, we have decided to implement the [ Data Cloud ] and this was really for us the foundation of the change because in less than 3 months, we have been able to put all our data in 1 layer and to be able to run reports.

So that was more or less, what, 3, 4 years ago?

2.5.

2.5. So they were stalled a little bit of a boring usage of Salesforce. They were a pretty nice-sized company. They were -- sorry, a customer. They were in the 8-digit business. It was good, 42 instances of Salesforce, but it was a mess. It was a little bit of a mess. And we were not growing with them. In fact, we were having conversations of even reducing the fear attrition that some customers want to use less. That's where we were. And then you implemented Data Cloud and you killed it. It was...

Data Cloud 3 months bring us new vision of our business, capacity to implement new processes to better serve our largest customer, which represents 70% of our turnover. So very important moment.

Okay. So then that sort of -- the way I looked at it is that really appeals to you. You have now a global view of all your orgs. It sort of stopped the bleeding, okay? But then you came to the next Dreamforce and then you saw Agentforce. And I really would love for you to tell this audience with the same passion that you told me yesterday, what are you doing with Agentforce? What agents are you deploying, the volumes -- the volumes that you are deploying them? And most importantly, what is the business value that you are driving?

Okay. I may be a bit wrong. But I will do it in English. Even if Alexa asked me to do it in French, I will keep English. We have started in U.K. in May '25, 6 weeks to develop our first agent. It was a prescreening agent. Today, we have 7 agents running in 12 countries, representing 60% of our top line. In U.K., our first agent was what we call a screening agent. It means we need to prequalify candidates before a real interview with our recruiters.

Year-to-date, we had 2.7 million conversation with these agents worldwide. And for us, the key business impact is a reduction of time to present the candidate to our customer. We reduced it by 40% by the usage of an agent. And we were with Parker a few months ago in a branch in the suburb of Paris, and Parker kindly visited the branch and had a discussion with our recruiters. And one of our lady as a recruiter talked to Parker.

This agent has changed my life because at 10 to 6 in the evening before leaving the office, I launched all my agents. They do the job during the night. And when I come back in the morning, instead of having to call 200 candidates, the agent has already prequalified 20 candidates. And one of the key advantage of this agent is that it is running 7 days a week, 24 hours a day. When our branches are closed, we are still doing business, prequalifying candidates.

In fact, a big percentage of that work for the agent is in after hours, right? Okay. So that's the prescreening agent.

We have also the interviewer agent, which makes real interview with voice.

You have an agent, an AI agent that interviews candidate?

Yes. Including voice. We are using Voiceforce, last name of it. 20% of our calls are currently done by voice, and we do 20,000 interviews per week with our interviewer agents. And this has a direct impact in terms of business. We increased what we call in the hiring area fill rate. It means our capacity to answer to the orders of our customer by 10%, okay, by using this...

You placed 10% more candidates?

For the orders when we put in...

With the combined human and agentic workforce that you have put [indiscernible].

Exactly.

20,000 interviews per week.

And what is quite interesting to see is that the agents allow us to go faster. It takes us less time to present candidates to our customer, and we present more candidates to our customer. So it's speed and quantity we have.

Got it. So... That's -- we have the Agentforce with Data Cloud made a big difference. Then Agentforce, 7 agents already. You talked about 2 of them, global scale, big numbers. By the way, AWUs in the millions, okay? Thank you very much. Awesome. And what is happening now? Tell me what's happened this week that you made a decision.

This week...

Because we are in a world of AIforce.

I remember when we went out of the branch in Paris, Parker told me, it's nice your story with agents, but did you notice that this lady has to close 15 windows in Lightning before launching the agent. You remember Parker, you told me when we went out of the branch. And now we have made the decision to roll out Coworker to our 27,000 users. When we were with Parker in the branch, the lady was a bit stressed, Parker was there and everything and she missed the path to go to the right window to show Parker the results of the agent. You remember Parker.

And now she just has to type in 1 command in Coworker to do it. Coworker, you see I have some [indiscernible]. So I have launched projects, fail terms, succeed terms. It's the first time that I see such a smooth implementation. We, as tech people, we are almost no more involved.

We have... Was there an implementation or was it...

No, it's [ choices ]. We switch on and the user group together, they are rating their prompts. They roll out between themselves prompts to do performance management in terms of pipeline, how to do a search and match for certain orders. They have created their library of prompts. Now they share between countries the library of prompts.

So it's just incredible to see how we have changed the surface of managing our business. It means from a sales or delivery perspective, and it gives really energy and joy to our people to have a new way to interact with the system.

I love it. And the incredible thing we presented 3 AI -- new AIforce surfaces yesterday. Coworker is the most basic one because it's already there. It's living there. You just need to click and then it opens up and it's like a co-work environment within Lightning. And we were not sure how successful that's going to be -- that was going to be. But then we -- it took the market by storm.

Then we also announced Claudeforce and Slackforce. So this one, when I heard the story, the first thing I thought is great. We're going to have a great deal with them this quarter because they have to pay for it, right? Coworker, if you saw the SKUs, it's only available to the Max or the A1E. But this guy, white hair, very clever. He turned that on because he already bought A1E. So he's now enjoying getting the structured and unstructured data that sits on Data Cloud for the last 3 years. And now Coworker is really driving insights and helping everyone.

So I'm super excited. So we've driven a lot of value to you. I mean, 10% more placements. I mean that's your business. It's like pretty impressive. We've gone on a journey 3 years ago, big numbers, but declining. I think in the last 2 years, we've done a few very -- the most important thing is we drove a lot of value to you. We've been able to monetize part of the value. That's the trick of the AI. Our business with them has significantly scaled, okay? We cannot share all the numbers that when I say significantly, it's significantly bigger. And we are just getting started.

Yes, but we need to renegotiate the new contract. Okay. Don't forget.

He has the best contract possible, right? So he signed an AELA 7 months ago. The good news is he signed an AELA, which has unlimited credit for all these use cases. He has A1E. So he had -- he can deploy Coworker, he can deploy whatever he wants. That's the good news.

That's the good news.

What is the bad news?

We need to renegotiate the contract...

The bad news is the contract is a 2-year AELA. And after 2 years, we need to renegotiate that. They are such an incredible customer. And my commitment to him is, if you come here on stage with me, we'll take care of you.

But we're going to -- we're going -- we're going to share in the value that we are driving with you. Listen, they're awesome. Thank you so much, and looking forward to driving more value. And you have one more thing to say?

Yes. One more thing to say is that we are only at the first step of the agentification of the complaints. Today, we have agentified human processes. It means we have taken human processes, and we have identified them. Now we are close to open the second chapter of agentification, which is to redesign our own processes based on agent and then to see where we need human. This will be a big change also for us. And all the agentification, in fact, is a transformation lever for all of us.

All right. So now let me do something before I call the next guest on stage. I have this little thing here. Paul, okay, I'm doing this for you. I don't know if I can do this, but it's going to be difficult but I'm going to try to do this. And so I've known Paul for many years, too many. And he's an amazing friend, and I admire him as an executive, what he's done over the last 15 years since I've known him. Nothing short of incredible.

In fact, he's done something that nobody ever in the history of any company has ever done, okay? You know the numbers of Anthropic. And super thrilled to invite him on stage. Please join me, Chief Commercial Officer of Anthropic. Paul, awesome. Great to have you here. And so Anthropic is not just a great technology company. It's not just a great partner, that it's a high growth. It's a high growth -- so it's not only a high-growth customer of Salesforce, but also very sophisticated, are growing very rapidly. To the point that 4 months ago, one of your sales leaders, Paul, was in a public event -- and they -- and she actually explained nothing to do with that. We were not invited.

She explained how she was running her commercial machine, her commercial engine. And then there was a picture there that had Coworker in the center, obviously. But then it had a bunch of pieces from Salesforce. He had Salesforce, Sales Cloud, it had also Slack. It also had Fin. At the time, we didn't know what that meant. Now we know that Fin is also part of the family. But it was incredible, and it really inspired us to look at how -- I mean, because they were very successful.

They were growing a lot, the speed of their execution. So we went in, we double-clicked, we talked to your team, we talked to you. And then Patrick got involved, Alexa, we realized, why don't we just build this for every customer there. And that was the origin of Claudeforce, our partnership. We announced the product itself to Salesforce in Claude this week. And I just wanted to ask you the first question, Paul, what are you hearing from your customers? Why is it so cool for your customers, for all our customers?

This is one of those wonderful products that your customers kind of organically kind of create ask for almost pull it out of our hands in that we were inside, as you said, with -- I think there was Kate, who was basically doing that at the time. We were using Salesforce in Claude long before Salesforce in Claude was the product that it is today. And then a lot of our customers jointly were doing the same, and you started to do the same internally inside...

Even before Claudeforce that it was funky and we had to connect and it was...

Because intuitively, it just makes sense. It's like if I'm working in something like Claude and Coworker, then I want access to the data that's there, how do I put it in? How do I analyze it in the way that I want to analyze it? How do I bidirectionally read and write. So -- and use it in a very powerful way. And we -- as Marc calls and as you call it Claudeforce, as we call it Salesforce in Claude.

How do you basically produce all of the connectors and the skills to just remove any rough edges and just make that a seamless process. And it's how I run the business every single day, and it's how our individual AEs prep for meetings and do all of the things that expect them to be doing. And it just unlocks a tremendous amount of value. Like you've got decades of sunk value and like investment that goes into a Salesforce implementation. This exposes that kind of supercharges it and it allows me to intuitively use it, have a conversation with that data and use it well.

You, yourself -- this is a script, but you, yourself, have been using Salesforce for many years. And you went back for 4 or 5 years in the dark, you didn't know what you were doing. I wasn't able to use it. You came back to the light. Okay, it's amazing. So our companies use each other's technology quite a bit. We are a very proud customer of Anthropic. We use our engineers, 15,000 engineers, they use your code.

We use your model to power Slackbot, which is transforming Slack and our company. We also use a little bit in Agentforce and we use Claude [indiscernible] in Slack. And we use -- we power Coworker. You heard Pierre is over the moon with Coworker. At the end of the day, the reasoning engine is you. And -- but you're a great user of our technology. So can you give us a bit more detail on Slack, the pieces, Slack, Salesforce, how do you use all our pieces?

So I think we are probably -- I don't know, we might be neck and neck with Amazon, for example, but I think we're one of the most intense users of Slack in the world in terms of for our headcount, just how much we use it. It is kind of like the high [Audio Gap]

While you're here, I would encourage you. I had the most phenomenal experience Monday afternoon in Agentic City, and I have to go back because they weren't all. But go over, in addition to what you heard today, you can truly see from 20 different companies the way they're leveraging the use cases, the value that they're bringing -- that we're bringing to them leveraging our platform. So I encourage you to take a swing through and spend time. But most importantly, we appreciate you being here today, and we're looking forward to answering your questions after this section.

So last year, if you remember, we talked about the FY '30 financial framework. And there were 3 pillars to that. Growth, we started at $60 billion. We closed the Informatica transaction and increased it to $63 billion. We talked about operational excellence. Again, a focus on durable profitable growth. And we also talked about our important metric that we look at all the time is free cash flow. And I'm proud to report, and you heard Miguel talk about it earlier, we're on track. 12, 18 months, we'll negotiate.

But at the end of the day, what we talked about, we're seeing happen. We are on track for organic revenue growth reacceleration in the second half of FY '27 and beyond. And that's happening because of all the innovation that you've heard about. We're also investing. While we're continuing to focus on profitability, I think Q2, we were at 34.1% on track relative to our guide for the full year. But we're also investing to scale profitably and to focus on the long-term trajectory that we see. And I'll talk a little bit about the opportunity.

Importantly, the investments that we're making are translating into momentum. Alexa mentioned the opportunity. It is pretty amazing if you think about what Gartner has laid out as the incremental AI spend from 2025 to 2030, $1.1 trillion. In the slide that Alexa shared on monetization, we talked about the fact that knowledge workers is a brand-new category. And that we expect software spend to literally double by 2030, all within the framework that we outlined.

So again, another high level -- there is a TAM opportunity out there, and we believe we have the innovation to take advantage of it. Most importantly, we see that innovation already starting to drive our flywheel. So we've talked about our platform. We've talked about the great innovation that we have across the various layers. And more importantly, what we see, particularly with AIforce is the opportunity to unleash or untrap the value within our platform. You saw the excitement in all of the various demos.

I'm not going to share my [indiscernible] center because it does have real numbers looking out. So I won't share it. But ultimately, we really see the future opportunity with the innovation that we have in place today. So I'm going to click in a little bit on the proof points around that. And I'm going to start with the AI native companies, one just left the stage, and you saw how they're leveraging our platform to drive their business. Importantly, 9 out of the 10 top AI companies are building and betting their future on us. That's a huge responsibility, but it's a huge opportunity.

You saw it earlier, 5-plus clouds per customer seat growth growing 62% year-over-year and 100% Slack adoption. We had an experience yesterday that I and my team led with COOs and CFOs. And we actually had the CFO of one of these AI native companies come and speak with us. And she talked about the fact that her morning and her evening ends in Slack, very much like mine. You heard Paul talk about it today. But it's not just the AI companies where we see this happening. We're seeing agentic enterprise expansion in action across our diversified portfolio.

If you think about everyone from automotive to a manufacturing customer, professional services, what we see is that in 80% of our top ARR growth stories, AI shows up. AI is the driver. It's not the only thing, but it allows that expansion to happen. So I just want to drill into one example here, public sector. Here, you can see that with Agentforce, the customer enacted the premium upgrade motion. They took on more Core and Data Foundations. They have monetized 4 of the 6 pillars that Alexa spoke about.

And that's all showing up in a single customer. The results and expansion across our platform resulted in about a 4x increase in ARR for that particular customer over 2 years. And as you can see, there are many different paths to that monetization or that uplift.

One of the things I love listening to Pierre, he talked about that journey that he's on, right? And so what we see is other customers are adopting these AI optionality, we're seeing the monetization of that happen. We're looking at the top 100 AWUs customers, and we're seeing within 18 months of Agentforce's launch, we're seeing ARR uplift about 2x as to what it was.

If we go further into our customer base, we're also seeing ARR meaningfully grow. And again, going back to Pierre's journey, and I'll go back to a slide slightly formatted a bit differently from last time, but we talked about that journey. It is a multiyear process. Pierre talked about the fact that right now, he's adding agents on top. But similar to the transformation we're going on at Salesforce, they still haven't necessarily rethought about their processes relative to Agentifying. So the opportunity is there.

And over a multiple period of time, we see as customers become agentic enterprises, a 3x to 4x opportunity for that ARR uplift. And that's what we're excited about. Again, proof points of the flywheel continuing in motion. So I want to go a little further into our financial performance. Marc shared this slide at the keynote. We've been in the business of helping support customer success with tremendous innovation for over 27 years. Now that's been a combination of organic and inorganic investments.

Last year, I talked about $10 billion or so in R&D investments. You can see the velocity of speed in terms of what we're investing to ensure that we're taking advantage of the AI opportunity. We're also focused on responsible M&A. We closed Informatica in November of last year. And you'll remember, we talked about our responsible M&A framework and making that transaction accretive within 2 years. We're similar to what you saw us do with net new AOV, Informatica was accretive in 6 months. And that was along with the fact that we continue to grow the platform. So again, disciplined execution.

So, I want to drill into our biggest acquisition, and I'm going to say it very simply, Slack is on fire. 2.5x increase since acquisition in revenue. You see all the innovation from Slack Code to Slackbot in Channels to Slackbot Live, amazing innovation. And Slackbot, which is a true friend to me and many others, 150% usage quarter-on-quarter. And as we've said, all the leading AI companies run their business on Slack. Alexa demoed for you, she runs her business on Slack, and you've seen it work. So it is another key growth opportunity for us.

So this progress that we're seeing from an innovation standpoint to increase ARR, it's underpinned by our profitable growth framework, and it's accelerating. The growth drivers we've talked about, the incremental monetization opportunities Miguel and Alexa spoke of and some remain the same, multi-cloud motion, pricing and packaging. You saw from both Patrick and Alexa, the Upgrade Megacycle that we see happening given the innovation we have in hand. We've got a pretty diversified portfolio geographically, different lines of business, sizes of business, different industries.

And the innovation, as you've seen, is pretty incredible. And we're investing in the capacity to make our customers successful. So on the margin side, 1,500-plus basis improvement between '22 and '27, and it continues while we invest. We're looking at ways to reduce the cost to serve, which helps our gross margin. We're investing ourselves in our own transformation, Salesforce on Salesforce or as we call it Customer Zero, and we're being disciplined. We are funding our best opportunities, but not every opportunity.

And that's the continued rebalancing and reshaping of our portfolio that we're doing to support our profitable growth framework. And what does this all mean? It means we're fueling our future with free cash flow. We estimate that this year, it will be about $15 billion. It gives us the flexibility to accelerate our innovation. So I want to drill a little bit into that last pillar of that framework, capital allocation.

There are 2 key principles for us. 1 is responsible M&A and strategic investments and the other is robust capital returns, which I'm sure you as shareholders all care deeply about.

Here's the framework I talked about earlier. It's still in play. It's still how we think about the investments, the acquisitions that we do and those that we walk away from. And if we think about it of late, I want to go into a few key focus areas for us. Data, AI accelerators and a new vector for us, AI Labs. Things that we look at when we're evaluating those opportunities, tech and talent, adjacencies, scalability. In the tech and talent era, I'll call out Doti. This was an acquisition that came to us as an opportunity between our M&A team and the Slack team.

It was a critical component to some of the innovation that you're seeing in Slack. And all the way to the right, we've talked about it, Slack and Informatica, clearly giving us great tailwinds that are helping fuel our growth in our FY '30 framework. So what is AI Labs? AI Labs is a new area of focus for us. It's going to provide us with frontier AI capabilities. It's going to help us accelerate our AI product road map, and it's going to help us unlock new opportunities because as you all know, the pace of innovation, the pace of change is rapid.

We've got a pulse to the ground and AI Labs allows us to be nimble, quick and adaptable. Three examples that sit under that umbrella were Regrello or as we call it Agentforce operations. Fin, which we just welcomed earlier this month and Qualified. You all saw the demo in the keynote of Piper. If you go to our website right now, you'll see Piper. And all of this is grounded in us delivering customer success because as you heard from the customers and Miguel and Alexa, everyone is approaching this agentic enterprise journey differently.

So I want to just drill a little bit more into this ecosystem. It's how we're thinking about one of our CEO's key priorities is do we have the right strategic investments to be successful. So Salesforce Ventures, amazing investments. You can see them there on the wheel on the acceleration wheel. I talked about AI Labs building. We're investing internally in incubation opportunities via AI Labs. I talked about the acquisitions that we're making that sit under that. And oh, by the way, you've heard about our partnership today with Anthropic.

You heard about our partnership with NVIDIA yesterday on stage, and there are many more to come. I think most importantly, we see this innovation flywheel as critical to our success and it allows us to stay ahead of the innovation curve. The other component that I mentioned is capital return. And I think the biggest bet that we've made in FY '27 is on us. We've been buying shares for a long term, $60 billion cumulative, but we launched earlier this year and we'll complete in October, the largest accelerated share repurchase ever.

The return over 40%. The expected average price, $182 a share. The expected share count reduction, over 14%. And again, that brings value to all of you, our shareholders. That is in addition to the $3 billion in dividends that we paid out. So to conclude, the framework that we talked about last year remains intact. We're very confident with the playbooks that we've talked about, with the innovation, with the go-to-market strategy and our path to $63 billion. The confidence is there. The TAM is there.

So to conclude, so we can go to the Q&A. We believe we are winning at the Agentic Enterprise opportunity. We're defining the Agentic Enterprise. The execution is compounding our value. It's unleashing the trapped value. And as I said, the opportunity is ours, and we're confident that the innovation that we're bringing will continue to drive the flywheel. So thank you.

I want to thank the whole executive team. We've had -- we're, of course, saving the best for last. We've had so many requests. People want to see into the mind of Marc, and Marc is with us, and he has graciously offered to give us his time to handle solo Q&A from the audience.

How many people here were -- saw the keynote demo yesterday? Great. How many people didn't make the keynote? Just a couple. Somebody asked me that I'm supposed to say thank you to all the people who sold us the stock at $150. I don't know what that means exactly. But thank you to those people.

We appreciated the volume there. We appreciated the volume in the stock. Okay. So we have -- do we have folks assembled? Do we have someone running a microphone?

I'm happy to do any questions or take a little bit of time...

I want to start with Brad in the middle. Can we get a microphone to Brad, please? Brad, do you want to just speak into [ mic ]. Do you want to speak in my [indiscernible].

This is interesting. This is a little bit different. We're doing a little differently this year, Marc. Brad Zelnick, Deutsche Bank. Thank you again for having all of us. Another spectacular Dreamforce...

You'll get the hot swap.

Perfect. And the sneakers never disappoint. I wanted to ask about Slack and as well the evolution of the business. If we reflect back 5 years ago, it was a highly strategic acquisition that you made...

We're very well-received acquisition 5 years ago -- it was everybody support in the room for the acquisition 5 years ago. Thank you guys so much for that.

Ahead of its time. But owning the interaction layer where customers engage, where there's a lot of rich valuable data, we saw the opportunity -- some saw the opportunity. Fast forward to today, we've seen great demos of Slackforce. Alexa, in particular, absolutely crushed it. Patrick showed us Claudeforce as you get pulled where customers take you. You've always been customer first, built on trust. What does the future mix look like? How should we think about the stickiness and opportunities to monetize, whether the customer comes through your front door or someone else's? And does it even matter?

Yes. I think it's such a great question. I think a lot of people know, I'm really just back from 2 months in Europe. And when I was in Europe for the last 2 months, I literally was with hundreds of customers. And it was definitely an awakening for me in many points, but it really resulted in the keynote that you saw yesterday. #1 is this. I would say that -- something is happening over here to the right. I don't know what it is.

Okay. I think we're about done.

All right. Great.

Here we go. Much better.

The #1 thing is that we saw this incredible thing was that -- I mean, it was like really one specific seminal moment. I was with this customer in Lausanne, Switzerland. I don't know if you've been to Geneva, just outside of Geneva. It's a great area on the lake. With one of our very largest or maybe it is our largest customer in Europe. And it was a long exhausted meeting. It was about 3 hours. And we're kind of getting to the end of the meeting, and I was with Rami, who's the CEO of our Swiss business. And we had already released Claudeforce to our distribution organization and our operating unit leaders.

But what we didn't realize -- and I kind of want to get into this a little bit more, is that the platform itself was so powerful that they were going to start using it and building their own application to run their business. And Rami was in front of the CEO of this company and started demonstrating his version of Claudeforce that he had given in a name. He had built this highly customized application to run Salesforce Switzerland. And I will never forget this, but the customer's jaw literally dropped open when they saw what was happening. Which was that the platform was able to kind of read across our entire database, all of our applications and build this highly customized application that let Rami run his business.

And this was kind of an amazing moment. It was only complemented by my own personal experience was we were in the focus groups for the Dreamforce keynote that you saw yesterday and Miguel, who's right here, we were in Beverly Hills. And -- we're going through -- we have customers there. We're presenting to them. I'm sure you know what we do.

It's very exciting. And then we're trying to listen to them. And I just kind of glance over and I look at Miguel. And I see that he has taken Claudeforce, and he is running an application that looked unlike anything I had ever seen. It was incredible, graphical, dynamic, very interactive, very intelligent. And he was kind of switching between that and Slack and so forth.

And it just all of a sudden occurred to me, everything is really changing because of what is happening at the interface lever. And at this interface layer, we're really seeing this kind of transformational moment. So this was that thought that -- and I don't know if I'm going to be able to express this correctly, but there are certain moments in our industry where everything is changing. And this is that moment for enterprise. We saw it when we were doing IBM mainframes and then we went into minicomputers, but then all of a sudden, we started to go into client server computing. We saw it when we were moving client server computing and then all of a sudden, we were doing cloud computing.

And now we're moving from cloud computing, and we're moving into this new platform. And that's why yesterday, I kind of went through these four strategic layers of how our software is architected. And for those of you who have been coming to these Dreamforces and going through these things, you'll see we really haven't had to kind of spell it out quite like that. The data layer where the data is integrated and federated and harmonized the application and semantic layer, which before we were just talking about the applications and the power of the applications. But now we're talking about how the applications and the analytics provide the intelligence for the AI, the agentic layer, which we have been talking about Agentforce, but not as part of the comprehensive platform and now a transformational interface.

And that transformational interface, I think, is why -- I think for a lot of customers, as you kind of travel around, they'll say, they really haven't had their enterprise AI transformation. I don't think most companies feel and probably a lot of companies are represented in the room that when you go to work, you feel like, oh, yes, AI has really changed everything in our company. Maybe people are still having their ChatGPT moment at home or on their phones. But when they get to their office, are they feeling like, yes, wow, everything is different in our company. But it's extremely clear to me, and I'm very confident that we are now at that moment where all of a sudden, we are going to see a rapid transformation of the enterprise.

And what we're doing with all of these customers today is showing them how to kind of guide that transformation. Now we're showing it to them with three different interfaces. One, we're showing it to them with all built on our Agentforce framework. We're showing it to them on Claude and Coworker. We're showing it to them with Slack, like you mentioned, and you've seen now Slackforce. And we're also showing to them with Coworker. In all three cases, you can see how the platform is intelligent, it's adaptive, it builds these composite applications, but also that it's very kind of like it's alive or it's a living interface. Only in the movies have we really seen software that looks like this.

We have never really seen software in the enterprise that has these characteristics. So this is definitely that moment. So for these companies and customers who see this experience, they all realize this is now the North Star. Now while we showed you three of these interfaces, okay, by the time we get back here or let's say, by even as we kind of get into January 1 of next year, I'm confident you'll have a lot of different interfaces. The interface, customers will be able to choose their religion up here. We're agnostic. That's, I think, one big advantage we're going to have is the agnostic aspect of the interface. But we're going to give customer choice. I think customers will have different religious preferences up here. There's no question.

And by the time we get back here to Dreamforce next year, I think we will see a lot of customers who have made a huge transformation. One of the reasons why it's going to go so fast is because those customers, as they start to deploy this interface, the interface is so intelligent. It's helping those customers implement it with them. So Miguel or Rami, they didn't have huge tech teams with them to implement this technology. They're not -- it's not like the -- it's not some type of incredible enterprise transformation. You're going to end up with an enterprise transformation, but the interface itself is so smart, it can do three critical things.

One, it can help administer Salesforce. That's extremely important because there's thousands of different characteristics of Salesforce. That's why these administrators can get in there and they have to kind of work with it in the different administration centers. Two, it's going to let them build these applications. And three, it's going to then let them operate the applications. So that's why you can get an operating unit leader or a business leader like Miguel or Rami to all that sudden have this incredible capability. So this is what we are really excited about. And I think we are really uniquely positioned to offer this to customers for like one really important reason.

And I think that you can kind of see how it kind of played out in that we -- I think I did a tweet earlier this year about Salesforce headless. And all of a sudden, this headless idea went super viral. But while that happened, I was actually kind of shocked because it seemed to me Salesforce has always been headless. People were excited that we had an MCP API, but we had an XML API, a SOAP API, a REST API. It's just a maturation of our API platform. And the other key characteristic beside that we have -- we're built on a robust API strategy, and we always have, is that our applications especially the ones that are built in Lightning. Lightning is not just a runtime environment, it's a design environment.

So those customers have designed those applications and have decomposed them into the metadata. The metadata is stored for all those customers in the database. So to your point before, that metadata then is recomposed into browsers today. Into HTML browser, Safari, Chrome, Mozilla, whatever. We don't have a religious choice up there either. Now, the browser level is transformed. And instead of having that level, you have Coworker or you're going to have other opportunities up here, quite a few, I'm sure, that let you build and chat and create and so forth. And now that kind of very -- call it, an intelligent browser. It can read through that same metadata and then recompose those applications.

And because our platform knows what the applications look like, what the data is, the user models, the sharing models, it is informed by the semantic layer as well. Instantly, you get that application. That is something I've never seen before in my career. I don't think anybody has ever seen this. I think it's by far the most exciting thing I've ever seen. It's totally unexpected and the value that customers are going to be able to receive will be incredible. And that's why we're -- that's why we have not -- I couldn't have been more excited to do the presentation yesterday. I think it was like, wow, this is really -- this is a dramatic moment, and we are uniquely positioned for that moment.

And I'm confident that -- and I don't know how many of your companies today use Salesforce to run your various operations. So there's a few. That is an immediate opportunity for you to turn on. It is not something that you're going to wait for, and you will be living in that environment the way Miguel or Rami are living in that environment today and soon, thousands of customers.

Okay. Fantastic. Let's stay in the front row and go to Kirk.

Kirk Materne, Evercore. Marc, to your point on religious preferences, you guys are agnostic. You had two of the biggest CEOs of AI companies with you yesterday. And I think part of the reason they look to Salesforce...

One, we had also Sam Altman here as well. So there were three. I don't know which two you were saying.

I was including the AI. I wasn't leaving Jensen out. So you had all three. But to the point on the AI models themselves, I think one of the reasons...

Okay, I won't tell Jensen he said that.

I'm not tall enough for him to make fun of me. So it's -- the point I was trying to get at was, I think whatever your model preference, enterprises, in particular, are looking for trust and they need to understand that whatever I do from a model perspective, I need my other enterprise vendors to make sure that my data is safe, the workflows I'm building are safe and are my workflows.

Do you think that you all can bring a level of trust to AI that will accelerate this? It seems in the enterprise market, not only have people been waiting for sort of faster time to value, but they're concerned about making a decision today that could backfire on them in a few years. What are you hearing from clients on that front right now? Because I do think it's important that they're watching models jump each other and they're saying, well, I don't want to make a big decision on something until I feel comfortable that the entire stack can move forward with me.

So I was just kind of curious if you talk about trust and it goes hand-in-hand with safety and then sort of what are clients talking about on that front?

Okay. So I think there's a lot of different questions in there, but I'll try to zero in on one, which is when I'm with customers -- like that customer that I was in Lausanne, I'll use that example. They're -- basically say this, they're not shopping around for models. What they say is we have three platforms that we run our business on. We have back office, SAP. We have front office, Salesforce, and we have a productivity suite, Microsoft. And the vast majority of customers that I met with, for example, in Europe, that is their core architecture. They're not kind of -- they're not -- I think we get brainwashed by these podcasts thinking that this is like the most important thing.

Well, this model now has this characteristic and that model has that characteristic. That's not where customers' minds are at. So customers' minds are like, they're operating their businesses, and they're using our technology to run their sales, their marketing, their service, close their books, provide capability to their employees. And that idea that customers are shopping around for different model vendors, I think that, that is not where their consciousness is. I think that the vast majority of customers will purchase their AI through packaged software. That's where the value will get provided. Through packaged software like ours, not in some bespoke, they're not -- they don't have the expertise or capability to do that.

That is you need to have a -- that is not where customers' heads are at. I think I can speak authoritatively on that at this point. And what I would say, though, in regards to trust, like when we hear other vendors say, well, if you're using such and such model, you're giving them all your intellectual property, that is not true. You're being told something to kind of reset your own frame on how to view that company. That is not true. That is a company that has their own agenda. And for the last 3 years, we've delivered a trust layer so that none of our customers' data has ever gone into a model and never will. We validated that and we have audited that. We provide a high level of trust.

We just reengineered the trust layer with Anthropic and OpenAI specifically, and it has 0 data retention. No customer data goes into that model ever. So when you're being told some of these things, you have to remember, people have their agendas that they're trying to get you to act in a certain way.

I think that, that was kind of at some level, the core of the SaaSpocalypse. And look, a lot of people made a lot of money on that. There's no question, okay? Well -- we're making some money on the other side of it, okay? But that we didn't create it, all right? And I mean I know who created it, how it was managed, but it's like that's in the same way other vendors have their specific agendas.

So when you're being told, well, you know that you're trading your intellectual property into a model, you need to realize that is not true. I don't know any example of that happening. So what I see is that what customers really want is they want an enterprise AI transformation. They can see that there is technology that could make them radically more productive and more successful. The way they are experiencing in their individual life where they can put some of their personal data into these models and then get some kind of an insight, they want that to run their business. So we know it's possible because we're living that now.

So, if you go and spend time with Miguel, you'll see, and you should take him aside or any of our executives aside, and you'll see that incredible experience. I'm confident we now have a highly differentiated, okay -- experience for the customers based on AI that is at the absolute pinnacle of what is possible. And I think we demonstrated it yesterday, and these customers have now an absolute true North Star of how to take their companies forward.

But in these companies who have spent billions of dollars of investing in Salesforce, like this example I give -- I'll keep going back and forth to this example in Switzerland. That company has spent billions of dollars putting Salesforce in. It just is going to go to an incredible new level for them. And the only reason that we know that is we've now done it for ourselves.

So I think we have more productivity and more capability because of what we've done with our platform, with the interface and also with Slack and other things that we're doing as customer zero, we're going to try to do this with every single customer. Does this make sense? I really wanted to directly address the ZDR issue. So I'm glad you set that up for me. Thank you.

Thank you, Kirk. Let's go to Alex in the front.

Alex Zukin with Wolfe Research. Truly inspiring presentation on all fronts. When we talk to customers and we do a lot of survey work, we do a lot of direct questioning, I think the main question is how much does it cost? And from the vantage point of you're putting out a truly differentiated amount of innovation and functionality, you're also being very flexible in how you price it, SKUs, consumption, value-based outcomes.

When you think about the distribution of how much value goes to the model versus goes to the harness or the app, in the future, over the course of the next few years, how much of that can you capture? And how do you explain that to customers of what they should be willing to pay for Salesforce?

Yes. I think that one of the most interesting things that's happening -- and this is a challenge for Miguel, so he should probably weigh in here and how I've tried to influence them, I'll tell you my own personal narrative, which is that every customer wants something slightly different on pricing. Some customers want per user pricing. It gives them predictability. They understand the cost structure. It makes sense to them. Some customers want per agent pricing. Some customers want consumption pricing. Some customers want usage pricing. Some customers want transaction outcome pricing, that is I've completed the transaction, therefore, I'm paying for that.

And some customers want business outcome pricing, which I have never seen before. But now they're like, if you save me so much money, I'll give you a percentage of that. If you make this much money, I'll give you a percentage of that. No one majority of customers can you drop in any one of these buckets. So what we've said to our entire sales organization is we are giving you ultimate flexibility to write the best deal for that customer. We want Miguel to know that he can go walk into a customer and say -- Miguel has a very large distribution organization. You know the size and scale of it. And basically not be constrained in any way and say, this is the right price for you. That is the way to get the most value.

And over and over again -- you have to remember, we have many different ways to get value. One is that agreement that we're writing. Another is also fundamentally reducing our attrition like we've done, also increases our value dramatically to our shareholders. So especially as we kind of cross now into the $50 billion in revenue, which very few software companies have ever done, we have a brand-new product line, and we have a brand-new approach to pricing.

So, you saw in the second quarter, attrition had reached kind of a very low level, contract length had come to a very high level. Cash flow and margin and revenue levels were very respectable, and we had very clear trajectory going forward where you saw the cRPO growth.

All of those things are critical and that I think our relevance has never been higher for these companies. Our ability to walk into any customer now and to show them a piece of technology that can radically transform their company very, very rapidly is unprecedented. And one more thing, it is not going to require a huge amount of people to do it because the technology is working part and parcel with the user to achieve the result. I only know this because it's happening in my company in a dramatic way. I'm not having to wait years for this transformation. I did not have to hire thousands of people to achieve it.

That is going to be very exciting for our customers. They still have to deploy Salesforce, but they can also do that in a much more expedited way. They still need to make sure their data is right because if their data is not right, their AI will not be right no matter what they're doing. And they can also deploy a fully integrated agentic environment as part of this. So when you get that working, it is just epic. And like I said, this is like stuff for the movies is what I'm seeing. And go back and look at the demos from the keynote, it's on YouTube, whatever. I have -- these were not canned demos.

Like Patrick, he has a lot of different versions of the demo because he has a lot of crazy prompts on how the demo was built, different style sheets and all kinds of different -- like he's laughing because he came up with so many crazy ways to present what he showed yesterday because he can -- he is building it. This is really special. And I think this is really going to transform computing. Do you want to now address the pricing issue? I don't know if I explained it. Did I answer your question?

Very good explanation. So Robin, in preparing for this event last year, you introduced a new mantra that we are using in every single conversation, which is we have to meet customers wherever they are in the journey. So we've come up -- at the time, we had 3 or 4 ways to monetize pricing commercial frameworks. We've added a bunch of them. Alexa and I, every time that we have a conversation with a customer, we find a new way to price our product. You'd be -- Marc, you've been happy and you'll be proud. I mean your message is very clear. We are now pushing very aggressively outcome-based pricing.

I mean Alexa and I meet hundreds of customers every quarter. Today, I met two customers. I made two outcome-based proposals. One was a PE firm. And what I agree with the PE firm is give me one of your companies -- portfolio companies that you're going to invest in and flip it after 5 years. We'll put an army of people and technology, and we get a percentage of the value that you drive. And at that point, let's say that the percentage is $300 million because they sell it, they buy it for $2 billion, they sell it for $5 billion. We get $300 million or $400 million ticket, and then we create a subscription out of that over 5 years. I mean we have multiple ways.

Adecco was on stage. They were reluctant to sign the first ELA, which took them to a different level of spend with us because they were not absolutely clear. Now not even a year into the ELA, we are now discussing, okay, this is going to -- we're going to run out of magic of how -- what are we going to do next? And I say, you know what, your time to place has been reduced by -- employees by 40%. Your number of placements have increased by 10%. What if we agree on a couple of metrics and then we share in your success? The guy wants to do it.

It's -- everybody wins here. And just to finish with one comment, Marc, I actually -- the reason I haven't -- I couldn't show my Claudeforce, I call it Miguelforce, but it's my Claudeforce is because I run the whole company data, and it's very obviously delicate. But you know what I did last night, I told Coworker that put a demo toggle on the top. So when I'm showing and I have it there and I already offer them to show my whole cockpit of the whole business of the company, I will click the toggle and then all of a sudden, all the numbers and the text gets blurred, but it's the same application. And it's -- we have made working with SaaS applications fun again. It's a lot of fun to put your hands in a keyboard and look at -- and then, by the way, I'm doing the more and I'm doing to the teams. It's -- yes, but we are doing business in the meantime.

There'll be translation available at the end of the program as well.

A Spanish trombone. Okay. Let's go to Karl back here.

[Technical Difficulty]

The first part is -- what I'm trying to do is what I've always done, which is actually provide a North Star for the whole industry. I'm trying to take the whole industry somewhere. We try to do that. Obviously, we want to do that with our philanthropic model. So let's take that off the table, the 1-1-1 model. We're trying to do that with our business model, but we are trying to do that with our technology model, too. And now we've slightly pivoted our technology model. So I kind of went through the four layers with you exactly like as I've done it now with hundreds of customers on this European trip and then that kind of got to the point in the keynote.

Now we get into two different places. One is I think a lot of you, but I don't know how many -- not all of the enterprise applications companies, #1, are API first the way we are; and #2 are not metadata first. There are still quite a few of these software companies that coded their applications. That is -- there is still kind of a ghost of client server past where the applications are fixed. They have not -- they don't resolve into full metadata. Some of them are full metadata systems, but some of them are not. So we have to kind of then bifurcate into those two worlds.

For the ones who do have metadata systems that are API first, you are going to be able to move very, very rapidly. For the ones that aren't, it's going to look a little bit like, did you see the keynote where we brought up the SAP screens and we're scraping and then we're working with the agent in the screens. That's a good example of what that's going to look like for companies that have kind of more of that fixed application set, okay?

A customer -- all companies are going to really have to move to what our architecture is going to have to -- what it is. The third piece is that when we're actually building these applications to help you run your business, we believe that the kind of things that the majority of employees in the company want to use, that is things that are focused on the customers, the products, the competitiveness of the company, its market position, its ability to collaborate, to share, to market itself, those applications are where we command and control the market.

So that is where we have a very unique and special position and that our applications are extremely relevant for our customers. In some cases, some of our competitors maybe have niche types of functionality. I don't have to get into all of the details, but some of their applications are not mainstream business data that is appropriate for large groups of users. Ours is. Because ours is for large groups of employees, that's where we're going to offer the most value in this application transformation. And I think that it kind of plays out with the customers. So I think you'll see it kind of go forward.

Obviously, we've seen a couple of killer apps so far in AI. One is the ChatGPT moment, okay? The second is the Claude Code moment where we have the coding agents. And now we have the third one, which is kind of the Coworker area. And Anthropic has been a really good idea. That's why we're so proud to be like investors in Anthropic. Like we -- John did a brilliant maneuver by buying hundreds of millions of dollars of Anthropic stock that has turned into probably what will be tens of billions of dollars of Anthropic stock. Obviously, we probably will not hold that stock in the long term. That's not our role. We'll end up probably selling it and paying off our ASR debt.

So it will be a good trade for us. We'll be trading our Anthropic stock for the 14% dilution that we gained back in our equity, not bad trade. I'll take it any day of the week. So thank you, John, for your leadership and great execution. But we, I think, are extremely well positioned. And I do think that, yes, we will motivate a radical transformation for the industry. We'll get there faster because we command and control the market because of the size and scale of our distribution organization and because we cut across small and medium businesses, large and very large businesses as well, and we can get to the -- we are getting to the market faster.

You can see that. I think that you'll probably say that maybe we're the first to show you what you saw yesterday. I think that, that's extremely important. I mean you follow the market super closely. And I guarantee you that by the end of the year, we will have several other extremely mission-critical motivating announcements at the interface layer. You can only imagine, I'm sure you could do the strategy at this point yourself on what all the different opportunities are at the interfaces.

Now that we have AIforce, we can slot in every other vendor. Obviously, we had to choose which vendor we were going to prioritize, which we did. But -- and I think we made the correct decision. We met the #1 AI with the #1 CRM. That was our goal. But we are -- there's no exclusive deals. We are going to provide a full family of interfaces based on whatever the customer's religion is.

Great. Let's go to Samik right here in the front.

Samik from JPMorgan. Great event. Maybe if I can get your thoughts on what's your vision with Koa, the model that you launched yesterday. You talked about training it on synthetic data, but when we think about sort of how does that drive value for the customer? How should I think about that? And clearly, with this event, your pace of innovation being ahead of your peers is visible. Does that sort of also then something we expect to see with how you train some of these internally developed models and take that forward in your portfolio?

Okay. Well, I'll have Rohan speak, and I don't know. Is Silvio here also or not? So I'll have Rohan then kind of fill in the details. But I think for a while, we've been -- we have -- you probably know we have a long history of delivering models about a decade. If you go to the Hugging Face, you'll see so many of our models have become very pioneering, including prompt engineering itself was built at Salesforce, but models like BLIP and xGen and others.

So we've always had a vision though of actually delivering a CRM model. And we think that there are certain things that we could do for customers with a CRM model that would be quite good, but we have not really wanted the price tag, okay? But with Nemotron, we now kind of have the ability to fine-tune a model without spending an exorbitant cost. And so that's the vision for Koa, is to take this vision that we have, which we call kind of CRMverse or Customerverse and apply it to the Koa model and to use Nemotron as the reference architecture. And I think it will be very exciting. Do you want to fill in the details?

No, I think, Marc, you captured it really well. So there's maybe just one more thing that I'll add. Obviously, there's 27 years of product making in CRM and sort of having those domain-specific models makes a lot of sense. And to Mark's point, like these open-weight models like Nemotron sort of make the post-training part of it a lot more amenable from a compute cost. That's what we've done. The other part essentially is this whole notion of data prep. So you can imagine like in the future, customers might want to use their own data and their own environment to sort of extend these open-weight models as well.

So there's an opportunity for us to actually build out the factory, like how do you go create that as a part of the data layer and make that available to the customers. Like they -- the most -- the bigger enterprise want to go build that on and they can do it themselves.

This is for a really discrete set of customers. As I said, I think most customers are going to want to receive the power of that through the applications themselves through packaged software. The vast majority of customers in the world will receive AI through packaged software companies. They will not be wanting to kind of spin up their own and running their own model. But for some discrete customers who want to have that kind of capability, we want to be able to offer that to them. Yes.

Wonderful. Thank you, Samik. Let's go to John.

It's John DiFucci from Guggenheim. So Marc, you and your team, I think, did a great job. I think everybody knows us at sort of exposing the SaaS-apocalypse as a hallucination. But I think that you'd agree with me that your stock is still pretty cheap. When I listen to your whole team today, one question kept coming up in my mind. Is Salesforce trying to become a next-gen CRM solution during the AI era? Or are you -- is this opportunity, this technology paradigm shift, an opportunity to become the trusted platform to bring AI across the entire enterprise? And I know you always think big. But I just -- is that something you aspire to?

Well, I think you have to prioritize that. I think you have to realize, #1, the most important thing for our company is we must be the #1 CRM. Nothing is more important than that. We have a tremendous position with our customers, as you know. Like I gave you the example of the three platforms. You could -- I think the number of customers that use that analogy with me was exhaustive. So we must maintain that position no matter what. And then number two is if we get there ahead of all the other companies, and we can show that through the flexibility and nimbleness of our platform, we can deliver this total enterprise transformation for you, we are -- we will do it. But it has to be a prioritization. You see, in software, if everything is important, nothing is important.

You have to choose. You have to decide what you really want. And I really believe the revenue opportunity for us the value opportunity, the differentiation opportunity and the ability to continue to lead our customers is, first and foremost, with that CRM opportunity that we have to continue to control that.

You saw -- I did -- I don't know if you saw this morning, I did a discussion with Roland Busch, the CEO of Siemens. He obviously, I visited with him also in his headquarters in Munich. It's a great example. Here's one of the very largest companies in the world, certainly in Europe, and they've been standardized on Salesforce. We must be their CRM standard. There cannot be -- that cannot be a discussion.

But do you think that there aren't other companies in there who would like to have that position? There are. Is it a constant discussion with them? Of course. We're in a highly competitive market. So we have to, #1, make sure that we secure that position. Now once we've secured that position, can we move on and then offer that capability to others? Absolutely.

Do we have that ability? We do. You can see it in the way we've architected our own systems. We are our own enterprise AI standard, right? But we have to be #1. And that's why when you look at the demonstrations that you saw yesterday, they are mostly focused on the customer area because I believe that is where we have to be #1. If you have a different position or think we should have a different strategy, I'm open to the discussion. I've definitely read a lot of what you've written recently. So I just think like -- this is the most important thing. And I think that -- were you surprised when you saw the keynote yesterday? Was there anything that shocked you? Or is it what you mostly expected?

I thought it was impressive, but it was -- I'd come to expect that kind of stuff. And by the way, your answer is exactly what I hoped you'd say.

Okay. All right. Well, there we go. Thank you. Well, that's what I believe. Okay.

Actually, right behind you, we have Adam, if you want to hand it.

It's Adam Wood from Morgan Stanley. So Marc, you talked a lot about the tech transformations to delivery of software that's happened. We've seen a lot less innovation on the payment side of how companies pay. Again, you alluded to how that's changing now. When we think about what replaces the seat is the unit value of software, do you have a vision of that in the midterm? Or what's the most likely model to replace the seat?

And then secondly, to the extent that's difficult to call today, how confident are you that the re-acceleration in the organic top line you see for the second half of this year can sustain smoothly rather than being kind of lumpy because of that risk of different pricing models proliferating?

I really am so optimistic because of the customer response that that's where I'm like, I think we got this really right. I think we are ahead of everyone. I think we reinforced everyone's confidence in the company and our ability to help lead all of these customers en masse I am shocked that we're there before anyone else. It is hugely surprising. I thought it was a shock that so many people actually push back on us even on the Claudeforce announcement. I don't think they really understand what's going on. I think there's still a lot of confusion. I think there's a lot of people who live in podcast world and don't live with the customers. I think it's a huge mistake.

I think what I see is we are so addicted to social media. I'm saying collectively, we're on X. We're on -- we're reading these things. We're watching these podcasts. We don't realize we're all part of a huge psyop and that misinformation is being shot at us constantly. And this is where the only way to break all that down is to get out with the customers. That is reality. When you're with the CEO and the CIO or the CRO or the COO or the head of sales and you are actually talking. What do you need to run your business? What do you need to buy from us? What is really important to you? What are my competitors saying? That is reality and these other things are not reality.

And I think this is where we can get really confused. And that's why for the last several years, I just kind of hang it up and I move in with the customers. And I think it's so important right now. And I think that, that's the way to really maximize the revenue for the company. I think it has been an amazing 3 years in terms of the way we've transformed the company. You guys have watched it. And the company that we have today is not the company that we had 3 years ago. Obviously, it's a different management team, but it's a different technology set. It's a fundamentally different position with the customers. And I think this is what the customers wanted.

We only did this because it's what the customers wanted us to do. In some case, they had an unarticulated need. They needed this, but they didn't know how to say it. They would say they wanted us to transform their company and make them AI first, but they did not know how. They could see the power in the models and the power and the capability, but they didn't know what could we do to make it happen. And then only when all of a sudden, we're like, let us show you this that they said, this is exactly what we want. And that is when we're like, now let's hit the gas on this.

So I was with one specific customer. I won't go through the details. They're a large automotive supplier, and they're based in Milan, Italy. And I was with them, and it was a fascinating conversation. They are huge customers of us on the front office, but they -- I was with the CEO and the COO, and they were doing supply chain scenario work using Coworker. And every morning, they're hand feeding the supply chain information into Coworker because that's the best they're doing. And then they switch over Lightning and Slack to run their business. And I'm like, I can show you a slightly different way to do this. They knew what they wanted, but they didn't know how to get it. It's not their world.

Just by making the -- they're here, but just by making that slight shift, then now they can go forward. And like I said, they can go forward very, very fast. So that is what is exciting to me. So I think we're going to see these incredible transformations. I think it's going to happen over the next 12, no more than 24 months. And this should be like an opportunity, as I've said for Miguel, we only had two goals in the keynote. One is to -- talked to the team, this was our collective intention. One, we want to motivate the enterprise transformation. Kind of to what John was saying, but a bigger -- on a bigger scale.

We want to motivate the enterprise transformation, CRM first, but across the board. Two, okay, we want to motivate the upgrade also. We also want to make it clear that we want them to step into the higher version to be able to get the full value. We know for customers who've already done that, they feel a huge amount of value by stepping into our higher version. This is where we're going strategically. That's why if you take apart our slides, you'll see that, that was our strategy all the way throughout. Does that answer your question? Okay.

Okay. So Marc, you're on a roll, but technically, we're at the end. What would you like...

All right. I could do one more question. How is that?

Let's do one more. Would you like to choose someone from the audience?

Not really.

Okay. How about I do that? Let's go to Tyler.

Tyler Radke from Citi. So you made the comment that you think enterprises are going to buy AI through packaged software. Obviously, we see the enormous growth that Anthropic, OpenAI are putting up. Given so much focus in recent weeks on trust, is it a strategic bet that makes sense for Salesforce to offer kind of AI reselling or some sort of model routing just as you're sort of this intermediary between the customer, their data and the model.

All of that is going to be built into everything you saw. The idea that -- we want to definitely be monetizing the token spend as part of our product line. I think a great example is -- I don't know if you've seen what we've done with Slack recently, but not only do you have Slack CRM where you can front-end Salesforce with Slack. Not only do you have Slackforce where you have the ability to build a whole new surface, but you have Slack Code. I don't know if we demonstrated to you today where you can code in Slack as a multiplayer experience. Today, coding agents are still mostly single player.

Slack Code is really the first multiplayer experience where all these companies who are using Slack already can now code in the channels. Our goal is to provide smart routing technology that we'll be able to monetize to be able to kind of sell you the tokens as well. But you can imagine that's an example of companies want to buy their AI through packaged software. So we should be able to like distribute those tokens through Slack. We should be able to distribute those tokens through all of our enterprise applications as well. Initially, it's not -- doesn't have to be step 1 for what we're doing. The most important thing is to deliver the base functionality that lets customers get the value. And then step 2 is let them kind of step into the ability for us to distribute the tokens. Does that make sense?

Well, I just want to thank you so much for coming to Dreamforce. I'm sure you know how grateful we are to all of you for being here. We -- I want to emphasize to you that the customers are untethered to you. We want you to go and talk to them. I know a lot of you do surveys and evaluations of what the customers' responses are. We don't want any constraints on any of the sessions that you can go to or trade shows or spending time with the customers. And I'm also looking forward to more quality time with everybody. Thank you very much.

Thank you so much. How about a big round of applause Marc. Thank you so much, Marc.

If we can remain seated for just a moment, I also just wanted to take a quick moment to say thank you to the IR team. First off, starting with Val right over here. You all know Val. Also, Alex, right in the back, everyone. Alex raising your hand. Thank you so much. Lauren is in the back. Here's Lauren back here. Sam is probably -- where is Sam? Sam is back here. This way. Is Anna in the room? There she is right there. Thank you, Anna. And everyone else that was related to this event in some way, I just want to thank everyone so kindly.

[Operator Instructions] So thank you so much for your time and attention.

Well, we're going to go to the keynote. Thank you guys so much. It's time for the keynote room with Marc Benioff and the Dreamforce main keynote, a new day for Trailblazers.

[Presentation]

And now our CEO and Chair, Mr. Marc Benioff.

Good morning, everybody. Is this microphone on? Good morning, everybody. We are so thrilled to have everybody here. What a great, great group this is. Wow, can you look around this room, it's amazing what's happening here. Totally incredible. We are going to have an amazing day. We're going to have an amazing conference. We have been on the road exhaustively for the last 2 months. I've been in Europe myself for the last 2 months. You're going to hear some stories about that. Then we got back here. United States, we did some incredible focus groups with all of you getting ready for this amazing Dreamforce keynote.

And let me just tell you this. We have never been more excited about a Dreamforce in our history. This is going to be absolutely incredible. You're going to see some truly amazing technology. You're going to meet some incredible people, and we hope to open a door for all of you. Now for the last 2 months, I've been in Europe. It's been amazing. I've met with hundreds of you. You've inspired me, you've energized me. You've really shown me what the future was. And it's been an incredible experience. But I'll tell you something amazing that occurred to me while I was in Europe. I had this awesome meeting with one of our top customers in Lausanne, Switzerland. And it was like exhausting long meeting. It was like 3 hours. And we were getting down to the end of the meeting, and I was like, I really want to show you what we're going to do at Dreamforce. It's going to be amazing.

And so finally, I had our CEO of Switzerland with us. He's here today. I said, all right, let's show them this amazing new technology, and we showed them the beginning of something we're going to show you today called AIforce. And they started using this platform. We've been using this kind of build and run our company now for several months. It's been transformational for us. In this case, we were using the version built on Claude, we call it Cowork. And this Claudeforce product, we're going to about to show you for the first time, the customer at the end of it said, this is really something like I see several things.

One, we had built a full new interface across their entire enterprise and did a complete AI transformation in real time with the customer. That was awesome. But number 2 was this. The customer said, wow, what I see happening right now is that you are really unleashing for me, trapped value in my system that I knew was there, locked in the metadata, locked in my systems and for the first time, for the first time, for the first time they saw that trap value coming into real value for their company. That's when we realized we could say, hey, I meet this #1 CRM Salesforce and unlock this trapped value from anywhere.

Now I want to just tell you, we are here to do many things with you. And one of the things you're going to see is that when we get into this incredible platform for our Trailblazers. Do we have any Trailblazers in the audience today? We've got a couple of them here. For our Trailblazers, this is the most empowering piece of technology we've seen. It really becomes a partner for you. It becomes a partner in helping you to execute the next generation of your enterprise. And these demonstrations that you're going to see are unlike anything we have ever demonstrated at Dreamforce before.

Now as you know, we are here to excite you. We are here to motivate you. We are here to inspire you but most of we are here to do what? Thank you. Thanks to each and every one of you for what you do for us every single day. We are so grateful to you for what you do for us. We know that you have a choice in vendors, and we know that you have a choice in where you have to put your time and to come here to this show and spend time with us, it's very meaningful to us. So thanks for being a great customer. Thank you for coming to San Francisco, and thank you for coming to this incredible Dreamforce. Now 20 years of these Trailblazers. It's been 20 years of these Trailblazers, 23 million Trailblazers all over the world, 90 countries, 700 community groups.

I saw Steve Mo got an amazing blue blazer. Where are you? Got your golden blazer, standup, Steve. Here's Steve. Great job. Yes. How long have you been with Trailblazer now?

I celebrated my Salesforce birthday, the day that my first user ID was created in August of 2003.

Steve, we could not be more proud of you and how you have lit up the whole community and you just guide us everywhere all over the world. Thank you for being such a great partner of Salesforce. Great job, Steve. Incredible. And Steve represents each of those 23 million Trailblazers all over the world. It's going to be an amazing show with all these Trailblazers. And as I said, I think the technology that you're going to see is going to really take all you guys to another level.

But listen, listen. This is the biggest and most important dream for us ever. And we've got all the greatest players here. We've got all the most incredible people from all over the world, the Dario is going to be in the keynote in a little bit. We've got Jensen is going to be here. We've got Sam speaking here later, Tony and Rob, we have all of our friends here. Isn't that amazing? And one more thing. We have 1,600 sessions for you. One really critical point about these sessions, we expect you to go to each and every one of them. We put a lot of work into them, 1,600. Let's do that. Okay. Tomorrow night, how was [indiscernible]? Was that awesome?

Yes. Well, we also have Usher coming as well. That's going to be awesome. And here's one more thing. We're going to be giving another incredible $10 million to the UCSF Children's Hospitals tomorrow night. You've now raised a total of $130 million through that concert every year for this hospital, amazing what you were doing. You already know, we so strongly believe that business is the greatest platform for change. You can see what we've done about $1 billion of grant since we started the company, 11 million hours of volunteerism. 65,000 nonprofits and NGOs on under service for free, and this is where I get to say this. If you are with a nonprofit or NGO and you are here at the conference, can you stand up and be recognized, please.

Amazing. And this is the opportunity for our commercial customers to realize that nonprofits and NGOs do not have as many resources as our commercial customers. Please support them and help them. And not only are we giving back at this scale, but one more thing. We are also deeply committed to San Francisco. Now with $156 million given for our public schools, $130 million to our hospitals, we have done 1.7 million hours of volunteerism here in San Francisco and Oakland, and we support 4,000 local NGOs. We've even given $13 million to our local police force to keep all of you safe. But we believe so strongly through this kind of $300 million in all-time giving that we have to give back to the communities that have given so much to us.

One more thing. We're also deeply committed to America as well. We now run 15 of the 15 cabinet agencies. 50 out of 50 states also now on Salesforce, all built on our core values, our core values of trust, the trust we have with all those Trailblazers. Their success is our highest priority. The innovation you'll see here at the show that we've been delivering day in and day out for 27 years. The equality of every Trailblazer, we so strongly believe in that. And the sustainability of our world, creating a net zero company. And that is why we've been influenced now as one of the most ethical companies in the world by [indiscernible] 7 team times, probably one of the most meaningful recognitions that we have received. Thank you so much, and it has delivered the largest and most trusted enterprise software company in the world.

An incredible slide as we slide into these very high ranks of these amazing companies. I guess you could say some people said, well, is Salesforce slowing? And I said, well, actually, I don't know if you noticed, we spent 3 years in the $30 billion in revenue. We spent just 2 years now in the $40 billion in revenue. And next year, we're entering into the $50 billion in revenue, looks pretty good to me, actually, the chart, not bad. Not a horrible chart. Thank you. And 27 years of amazing innovation inspired by all of you, especially our Trailblazers, incredible what you have been doing for us. And that innovation, it's organic, it's inorganic.

And let me also say this. It's also these incredible companies in San Francisco that we've kind of become part of our Ohana, and we want to welcome starting today, Fin, now part of Agentforce. The highest performing customer agent in the world, amazing company. Anybody here from Fin, stand up if you were Fin. What you be recognized we have our Fin team right over here. Congratulations, and welcome to Dreamforce. Great to see you and fantastic. Okay. Look, it's been 10 years of pioneering AI here in the enterprise. You have all going through this with us. It's been an amazing journey. There's no question about it.

And then we've been hearing all this kind of crazy nonsense about the SaaSpocalypse, especially over the last 6 months. It doesn't have been wild.

But I think that we realized that SaaSpocalypse, but it's not about the end of software, but it may be about the end of software that makes humans do all the work. And I think you're going to see that especially at the show. It's been incredible to watch the success of Salesforce this year, the growth, the capability of our customers. In our second quarter, we just delivered $46.4 billion in revenue guidance, delivered another $1 billion in free cash flow, 14% CRPO growth. It was awesome. Our lowest customer attrition ever, our longest customer contracts ever. And it's all driven by the deep commitment that we have at Salesforce to do this, to help you connect with your customers in a whole new way.

How do we do that? And now as you start to begin to see the technology that we're about to reveal, there's this fundamental question. And the question is this, how are we unlocking this trapped value across every single app? Now at Salesforce, we are starting to do this. It's been amazing. We already have 7,000 users on Claudeforce. So while you're at the show, take our Salesforce employees aside and say, show me how you're running Salesforce. You'll see the Slackbots that we have running and the coworkers that we have running in Lightning and that we generated $500 million last quarter in pipeline, incredible with our new Hunter and Piper agents, but we resolved 5 million service conversations with Casey on helpsalesforce.com.

We're being the change that we want to see with you. We believe that every company can become an Agentic Enterprise, and we're going to show you the best way to do it by doing it ourselves. It's been awesome for us. And as I've gone around the world, I see it in all of you. I see it in your eyes, your inspirations, your ideas. I see these companies changing and transforming in ways I could have never imagined. Now AI has opened a door that's never been opened before. How many of you in San Francisco while you've been here have already jumped into a Waymo? Anybody here jump in or Waymo? Not too many. Well, you might want to try it. It's kind of an amazing experience. It's an autonomous vehicle driven by AI.

And one of the things about it is it's kind of a metaphor in many ways. I told you about my trip to Europe, well, when I was in Europe for 2 months, I didn't see a single autonomous vehicle. That was kind of a metaphor to me. Well, we might be living at it here or if you go to Shanghai, maybe it's AI on steroids right now. But when I went to Europe, I didn't see that same kind of transformation, that same eye kind of same AI global, I would say, unbalanced capability, where some countries have more AI than others. This is a very real thing, but it's also true in our customers, too. When I go to visit all of you, and I just spent my time with hundreds of you, probably what I love doing more than anything else. But what I saw was a lot of unevenness what comes to AI.

So that idea is that when we start to look at what AI can do, yes, the door is opening, but not everyone has gone through it yet. In many cases, some of our customers are kind of stuck. They're kind of stuck in the old world. Maybe they're even using ChatGPT or Claude Cowork during the day at home. But when they get to their businesses and they get to their enterprises, they still don't see that enterprise transformation. Do you know what I'm saying? And I think we have to do more to help you accelerate that transformation. But one of the key parts about that is this, models alone cannot run the enterprise. Models alone are not going to show us what's going to -- what's possible. Models also are very probabilistic.

You know what that means, they're probabilistic, like they kind of know what's going on but they're not grounded in a single set of truth like your traditional Salesforce apps. I have a friend back home, and they've built this amazing app on Claude Cowork. And every time he shows some of the app he goes, well, this is this amazing app. But this number, I know it's wrong every time. I know it's not exactly the right number. I said, that's because it's not grounded in your single source of truth. It's not grounded in your core data set. And then how can you do that? The idea really is how can you bring that probabilistic world, the AI world, the deterministic world. The deterministic world is where all that corporate data is, where the single source of truth is.

That is the question that has been on our mind. And that is the bridge that we want to cross because when we look at enterprise intelligence, yes, we love these probabilistic AI systems. But how do we deliver the governance and the security and the workflows and the apps and the rules and the context to all of them? How do you bring all the power of the corporate system into the AI? And then one more thing. How do you then layer in these 4 critical layers? One, how do you deliver the data because we all know you've got to get your data right to get your AI right. Two, how do you take the power of the apps, which have all the intelligence of the business and all of the capability and the metadata and the semantics that deepen the analytics layers, the semantic layer. How do we unleash a digital workforce with agents like we've been talking about here for the last several years? And how can we build a radically different interface on all of those things?

These 4 layers data, apps and semantics, agents and interface, that is what we want to transform right now. And to make that happen to really bring enterprise intelligence to run on our core system, we have to unify the probabilistic intelligence with that deterministic intelligence. And the interface in agents remain probabilistic. And that is great as long as they are grounded inside the deterministic data, the source of truth of data, the data layer and the absence of semantics as well where you have spent so much time educating those applications on who your businesses are and how they run.

So at this show, what we will do and what we're going to do over the next 30 minutes is show you those 4 layers in an incredible new way. Data 360, a layer that helps you to integrate, to federate and harmonize all of your data. And that means connect everything in your enterprise together to smooth it out, okay, and then let it drive the AI. And two, to unleash through a new headless version of our applications, the fundamental application and semantic intelligence to drive that intelligence and then let the agents and let the interface do what it's going to do.

When we get to the interface layer, I promise you this. When we get to the interface layer, I promise you this, you will see an interface unlike anything you've ever seen before. It will transform your enterprise. And that is what is exciting. Now people been saying, as I've said that to them, how is that possible? How is it possible that so much could change so quickly at the interface layer? Well, for those of you who know how Salesforce operates, we operate through a metadata platform. When you're working in Lightning and you're running Lighting applications and where you're building on Salesforce, you're writing in metadata, you're writing in the objects and the interfaces and the rules. You're writing in the fields and the relationships, you're writing in the permissions and the business logic and security and analytics you're writing into the metadata and then it gets decomposed into the database.

It gets pushed down into our database. It's not fixed. I loved it when we announced [indiscernible] and everybody thought we somehow had cut the head off of our apps. There was never any head on the apps. It was always in the metadata layer. It's just hard to explain that to folks who are not Salesforce folks. So it's down in the metadata layer and then it's been recomposed into Safari or Chrome. It's been recomposed into the browsers or into iOS. And now it's going to get recomposed into these AI interfaces. It gets recomposed into Cowork. It gets recomposed into Slack. It gets recomposed into Lightning. So that is what becomes really dramatic. It's an interface revolution. But we've seen a lot of interface revolutions. We went from DOS to Gus. And we went from GUS to web, and we went from WAM to mobile.

And today, you're going to see, and I bet a lot of you have not yet seen this, you're going to see AI interfaces that are dynamic and intelligent, that are composable and alive and that can work with you in the fundamental administration of the applications, can work with you in building the applications and work with you in operating the applications because the interface itself is alive. Some people say the interface is living. It's really interesting. I'm going to let you determine what you think about that interface. Now to make all this happen, the first thing we had to do is write a layer that would empower all that. We knew we wanted the dynamism. We all know we wanted the intelligence. We knew that we wanted to keep go down into that metadata and pull it up. We knew that it had to be secure and governed.

And one more thing. We knew that it had to have zero data retention. Now 3 years ago, we were here, we talked about zero data retention. That means this. When you're working in Salesforce apps and when you're working on our platforms, your data is your data. It does not go in the models. We've audited it. We've tested it. We've tried it, regardless of what you've heard on media or from other vendors, let me assure you that Salesforce products are built with zero data retention, tested over and over again by our security teams. Your data is your data. You're not training any other model when you're using our products. This is extremely important, having ZTR built in. And then when you put it all together, when you put all those things together, you get AIforce.

You get those capabilities and you get this live interface that untraps the value. And then we're going to deliver for you AIforce in many forms, if you're a Cowork user, you're going to get Claudeforce. If you are a Slack user, you're going to get Slackforce. If you're a Lightning user, you're going to get the new Coworker. And there will be many others too, and we will also deliver an SDK, so you can build your own AIforce applications as well. This is really the most exciting thing I have ever seen for Salesforce. Because we have been talking about the data layer for a long time. We have been talking about the applications and semantic layer for a long time. And we have been talking about now the agentic layer for 3 years.

And now for the first time, we're talking about this, a live interface on Salesforce, unlocking this trapped value, finding the patterns across all your data, getting answers in seconds, not days and now anyone can get that value from Salesforce. That is the power of AIforce. And while you might not have seen this from other vendors yet. I assure you, you are about to see something that I'm sure that will get repeated over and over again throughout our whole industry. Just like the Trailblazers have sat in this conference now for almost 3 decades, I'm confident you will see this pattern repeat everywhere else.

We saw Steve Mo. We saw how incredible our Trailblazers are. They are going to be even more incredible with AIforce. They are going to be even more powerful to administer, develop and operate these systems across all of these layers. It is really awesome. And you're going to see the demo and pay attention, it's going to be really incredible. Now we've talked about the vision of lighting up the data, making sure your data is ready for you. And two, making sure all of our apps, which were always built this way, always built with metadata first so that this moment would be a moment they could take advantage of are also ready for you; and three, the agentic layer as well so that it can really deliver that value.

Now I know many of you are already starting to use these incredible agents. How many of you have been on salesforce.com or informatica.com or already using Piper, by the way, this qualification agent, quite a few hands. It is amazing you're talking to Piper, we're going to demo it in a second. It's helping to qualify your prospects and your customers. Hunter went out and generated $500 million of pipeline for me last quarter it was kind of amazing what Hunter did. Casey answered 5 million customer service issues for me so far since I launched about 18 months ago. And Paige is doing our ITSM work and HR capabilities. And Marshall is doing this incredible work as well in the supply chain. And now we have Fin, which is the #1 customer service agent. We've been using that on help.salesforce.com already. It's incredible, Fin is. And we got that running, I think, in about 12 days at Salesforce. It's an incredible capability that you can go super fast with.

That idea that you have those 3 layers, data, absent semantics and agents and then add the interface on top. Adding the interface on top means that everything is going to change. So one of the things I noticed when I was with all the customers for the last 2 months, and it was pretty incredible being there with everybody. But the applications kind of looked mostly the same that they've looked like for the last few years. And I would say to the CEOs and CIOs and CEOs of these companies, do you feel like you've gone through an enterprise transformation? Well, we've got this pilot going or we tried this or we did that? I think when you put this interface on, everyone will see that their systems have now changed. Incredible.

Okay. So let's take a look at Quadforce. This is Salesforce and Claude. It's the #1 AI meets the #1 CRM. It is that live interface built around you, if you are a Cowork user. You are going to love this platform. We use it at Salesforce. We use it aggressively. We turned it over to our operating unit leaders several months ago. They have been running their businesses. When we were doing the focus group for this room, I was sitting next to Miguel Milano, our new Chief Operating Officer -- Miguel, are you in the room? Well, standup. Here's Miguel Milano, our new COO. Congratulations, Miguel. I was sitting next to Miguel and he was using Claudeforce to run Salesforce. It was incredible. And he was also using Slack at the same time. It was hard to believe.

Now for those of you who use Slack, how many Slack users do we have here, raise your heads? Quite a few. Oh, I'm surprised. Well, so Slack is where AI works. It's pretty awesome. I mean we bought this company 6 years ago. We have invested a ton in this company transformed it. It's done amazing is to become our very fastest-growing product. So many AI companies here in the Valley like within a mile or 2 of this building, all run on Slack. It's awesome. But one of the things to about Slack now that you have is Slackforce. So that same user interface that is running from AIforce in Claudeforce is also now running in Slack and you're going to see that at the show. And for those of you who are coders or who are using coding agents in your businesses, did you know now you can code inside Slack?

Coding has traditionally been a single player sort board. It's a single player sport, but now it becomes a multiplayer sport, and you can write review and ship code together with agents and all of Salesforce is available inside Slack with Slack CRM. So you can use Slack to run Salesforce. And many of our customers are already using Slackbot millions every single week where they're reading across their DMs, they're reading across their channels, they're reading across the whole data set and letting Slackbot provide the AI for the enterprise. That is why Slack has become Salesforce's fastest-growing product.

Okay. And there's one more piece of this. For those of you who are on Lightning, anybody here use Lightning by any chance? Okay, we have a few Lightning users out there. Now Agentforce Coworker is running inside Lightning as well. So now you can have the same live interface in Lighting. So if you use Cowork, you've got it. If you have Slack, you've got it, if you've got Lighting, you've got it. And I bet by the time we get back here a year from now, we'll have dozens of live interfaces running on Salesforce. We have transformed our architecture. Look at this. You can see that core data layer with Informatica and MuleSoft. You can see it with Tableau and how it's working to deliver that core system.

And then you can see the application layer and all the different applications are now headless and providing that API and that capability. And the agentic layer as well and all the agents that you need to run your business or build your own agents with Agentforce also. All of these incredible pieces helping to create and empower those trailblazers that are so important to us. We want to make sure that Trailblazers are successful. They've built 12.8 million custom enterprise apps. They've done countless Apex calls and [indiscernible] calls. They've built these incredible systems. It's been awesome what the Trailblazers have done and this is going to elevate them to an incredible new place. And we're delivering new versions of Salesforce, including our new AIforce Max Edition, which bundles all of this together so you can have one core product that you can acquire that delivers all of this functionality to you.

One product, AIforce Max with all of our functionality built together. So there is no concern. You not have everything you need to be totally successful right now. So please welcome to give this incredible demonstration, our new President of Applications and Marketing, Patrick Stokes.

All right. All right. Thank you, Marc. Good morning, Dreamforce. Facing back with all of you, I love doing this every year. I hope I get to keep doing it. We'll see here in a moment. So listen, 6 months ago, we held an event. It was actually right across the street to launch Slackbot. And at that event, our Co-Founder, Parker Harris, you all know Parker, he's sitting right over here. He looks right into the camera and he said something a little weird. You want to know what he said? I've got the clip. Let's actually just watch the clip.

[Presentation]

Crazy thing to hear, right? Crazy. If you're any other software company, that would be a terrifying thing to say because other software companies, they think their product is the UI. But that's not what we think at Salesforce. That Salesforce, our product is the trust that all of you, our customers put in us to hold your data, to hold your workflows, your business processes, your permissions, your security rules, that is the Salesforce product. And we asked ourselves, what would it look like if we could plug that product in to a new interface, to an entirely new interface agentic interface. And that's exactly what we set out to do with our friends from Anthropic.

Now here, I am inside of Claude Cowork. And I'm about to show you my interface. Everything you are about to see is something that I ask Claude to do for me. And before I show you, I'm going to give you one last warning, which is, despite what Parker just said, you are going to see just a tiny a little bit of Lightning in here. Bear with me, okay? Here we go.

Here are your deal signals for today. Okay. body, this gets in. The quarter sits $8.2 million short of your $285 million target with 3 weeks left to run. NorthStar needs your attention, $4.8 million, 10 days past close and a block security review. Those are your signals pal, your command center has the answers. Trailblazers. Welcome to Dreamforce, digital Parker out.

All right. Thank you, Parker. I appreciate the insights. So look, this is my Salesforce. Like I said, everything that you see here is something that I specifically asked for and Claude helped me bring it together. I can see if I expand this, I can see my pipeline coming in from around the world, everywhere I've got a new opportunity opening up. I can see all of my service data coming in where my cases are landing around the world, so I can see where my hot cuts are. I can even see all of my marketing data. People interacting with all of my different websites, interacting with Piper on my website, signing up for new campaigns, interacting with campaigns all in this live interface.

But this is more than just an interface to look at data. Remember, we're combining this with AI. In this case, Parker, who's got a lot of insights, right? And Parker reminded me right when we started there, then I have a problem with NorthStar communities. And he said, you should probably take a peek at that. Now I know NorthStar happens to be based in San Francisco. So I can click into San Francisco here to drill into the pipeline that I have. I've got 2 accounts that have opportunities. I'm going to click into my NorthStar account here. And there it is, there is that security review that is blocked. And it's actually kind of bad. We've got $4.8 million in open pipeline that is not going to close unless we get this thing moving.

Now I wanted a way to let my team know that this is unacceptable. And I did that in the most Patrick way that I could think of, which is I ask Claude to code in a sad trombone. And that actually sends a real activity over to Salesforce. So that logs in this account record, so my team knows that I'm pissed off. If I wasn't pissed off, I could do something that I've always wanted to do, which is blow an airhorn on the Salesforce stage. I can do this as many times as I want. Nobody can stop me. Okay. But to get real for a second, what do I actually want to do? What I want to do is I want to send a Slack message over to my team. I need to get them rock and roll in on this opportunity. So we're going to click into the Slack interface, and I said, hey, give me a way to have Claude generate, but let me be able to rewrite the message if I want to, so I can do that right here.

We can say, hey, Dreamforce, it helps if you spell it right, although it actually doesn't really matter, God, it's hard typing on stage. There we go. And then all I have to do is post this over to Slack. And if I get over there fast enough, there it was, you saw it just pop up. 1046, there is our message right over to Slack getting our team rocking and rolling. Okay. Now I bet you're probably wondering how this all actually works. And let me tell you. So it all starts right over here if we go into, excuse me, plug-ins you'll see we've got this brand new logo Salesforce for Claude. Now this plug-in solves all of the hard problems. Who's heard of MCP servers here? Who knows what MCP servers are and how to use them? A lot of hands are going to go down, right? They're hard. This is a new technology. Who knows how to enable zero data retention, right? Not many people who's written skills for Claude, probably a few more of you right?

This is hard stuff. And we sat down with our friends at Anthropic and said, what if we could put all that together into one package, one plug in that if you have Claude Enterprise, you can just go in, turn it on and pick the users that you want to give access to. So that is exactly what we did. But what about that UI? How did I create that UI? Some PM didn't create that for me somewhere. That is a UI that I personally created for myself by just asking, and I'll show you that conversation. Here is the complete conversation. You can see where I started. You get up and running on these interfaces. I'm not kidding in like 6 to 8 minutes. And then you have a few days worth of back and forth as you iterate and you add all of the silly features like I did with that trombones, and that awesome 3D Parker that you saw. All of that, look at literally create avatar, Parker Harrison is Lightning suit. I to give it some things. But that is how we do it.

Now the last question I want to answer for you is how on earth do you get this thing, right? Well, we can show you. Now Marc mentioned the pricing and packaging earlier, but this is Dreamforce, baby. We don't want you to worry about pricing and packaging. What we want you to do is go get your hands on this thing. So for the time being, everything that you just saw is in open beta. Any one of you can go, you jump over to the AppExchange. I'll show it to you right here, right, again, inside of Claude, look for this plug-in within the op exchange, fill out the form, give us your org ID and we will turn you on today. If you are not here with us at Dreamforce, of course, we've got a Trailhead. I blew right through the applause, but that's okay. You'll do it afterwards. And if you are here with us at Dreamforce, I beg you get over to Moscone West, it's across the street, we have AI learning labs, you will get in there. There is an ocean of laptops with Claudeforce setup, you can learn how to build your own command center and walk out of here with your org turned on. So that is AIforce.

All right. Thank you. So now we have something incredibly special. And to do that, I'm going to reintroduce Mr. Marc Benioff. Marc?

All right. Great job. How about a hand for Patrick? Who's excited about that interface? That is just the beginning of the interface that you're going to see. I think AIforce remains the most exciting thing we have ever done at Salesforce and to see this first version of it running on Cowork and Claude, which is amazing. I don't know if you saw the cover of Time Magazine today, but Cowork is also on the cover of Time Magazine. And here we have the CEO of Anthropic right here, Dario, hello, how are you? Great to have you.

Fantastic. Great job, Dario. You're doing an amazing job leading the AI industry, guiding it all right from our Slack headquarters building. We watch you every day -- it's all good karma, isn't it? And we're watching you doing an incredible job. You've become the #1 AI in the world, 3x larger than the nearest provider. It's been amazing to be partnered with you. It's all been very karmic and kismet and good. And one of the things I have to ask you before we get into the interview, I want to talk to you about AIforce, I want to talk to you about Claudeforce. I want to talk to you about how you're running your own business. But wow, you -- I mean, everything from your first essay that you wrote about Loving Grace to the essay we did on pacing has been incredible. But give us your state of the AI industry and pacing and where your mind is right now?

Yes. I think it's maybe easiest to explain it by an analogy. So let's say you're running a car company and another car company, not yours, they have some kind of safety incident, right? Something goes wrong with the brakes, something with the manufacturing. Now you believe you have the best safety record in the industry. What do you do? How do you approach it? Obviously, it's very tempting to attack your competitor and say, these guys are on, say, for safe. But I think the more responsible way to respond to it is to say, hey, first of all, let's look at our own record, right? We may not have had this big high-profile incident, but I'm sure we're not perfect. Let's look at everything and we can always be better. Let's make our practices better.

Let's recommit transparency, let's invest more in safety then let's organize the rest of the industry and say, what can we do to set standards for everyone? What can we do to make everyone's standards better? And then finally, there should be an international component to it. And that's essentially the 3-step plan that I laid out in the SA, the first of which we committed to, the others of which we're going to have a dialogue with the rest of the industry about. And I think that's the way to lead the industry forward to set an example to say that everyone can always be better.

Let me ask you a question, Dario. It's been amazing what you have been able to accomplish, like I said. But there are a lot of surprises also that you have gone through. So just tell us what has been your biggest surprise over the last decade? And you've obviously been heads down in AI now for 10 years, really in earnest. What's your biggest shock about where we are right now?

It's been the pace of the progress, not just of the technology but of the economic side of it. So we -- I formulated the scaling laws along with some of my cofounders, which tell you that AI models get smarter and smarter, the more compute grow into them. And so we had a prediction of just how capable these models would be. But I think what we didn't appreciate is that it would lead to these companies growing so fast to all these incredible products that we're working together on and you've been talking about today and how quickly they would become central to everything that's going on in the world, right? I just watched the talks before that, where you guys talked about MCP and Cowork and Claude skills, and these are all new things that people are just beginning to use, but they're being adopted so quickly because the economic value is so great. So I've just been surprised at such a pace as possible in the business world.

Well, I'll tell you at Salesforce, we are huge Cowork fans. And also when I was in Europe, I met with so many customers who are Cowork fans, it's incredible what you've done with the product. I am noticing over your shoulder here, you've got a few Salesforce alumni behind you, helping you to run your company, which...

You train them well, Marc. You've trained them well.

Trying to do my best to help the whole industry. I measure myself by how many CEOs, COOs and CROs are in other companies that all came from Salesforce.

You do a fantastic job of that. No, truly. Truly, you do.

Thank you. But what I want to say is this. I want to say, yes, we have to keep our eye on the pace of AI. It's been amazing what's happened over 10 years. No one could have written the story. Now these interfaces are truly amazing. The demo we just saw is remarkable. I think every customer here wants to have that same interface on all of their systems. Everybody wants to unlock the trapped value. What is your kind of recommendation for everyone in this room and the virtual room? Obviously, you can see we have 10,000 people all over this room, but 10 million joining us online. What actions do you want us to take?

Yes. So I would just say, when I look at the technology versus how much it has diffused even if we freeze the technology in place, which we're not doing, pacing the frontier is not, not doing that at all, the technology will keep getting better. But it just suppose we freezed it in place. Even if that were the case, we're making use of maybe only 5% or 10% of what the possible value of the technology is.

So I'll give you just an example that relates to what we're doing together and to the CRO, you just pointed at, which is that Paul, every day now, he wakes up and if he wants to look at our head go-to-market, prospects, he still uses Salesforce but he uses Salesforce with Claude. He talks to Claude using Salesforce's data about what are the are 10 biggest deals we're doing this week. What are the 10 deals we're most likely to lose? What are the themes in those deals? How do we win those deals? Just making use of all the information fluently with Quad. And it's really incredible and the fraction of people who have seen this and seeing that this is possible is still small compared to those who could gain utility from it. And so there is still so much diffusion left to do.

Well, we are really excited to wake everybody up at the show to show people what the possibilities are for the future, how we're going to show them that data layer, which, as you know, I think, is really critical to having that grounding of the data. We've talked about that so many times. The application of semantic layer giving it the AI, even more intelligence capability, the agentic layer that you really have helped pioneer and now the incredible interface layer, and we're so excited that the first one is on Cowork. Congratulations. Well done. You're doing a great job. Thanks for everything you're doing for the whole industry. We're so grateful to you.

Thank you for having me.

Really thrilled. Thank you so much. All right. Thank you for coming to Dreamforce. Please thank the Anthropic team. And now please welcome back from that brief intermission, Patrick Stokes. Give my hand.

All right. Me again, just for a minute. Okay. So I think you can see how special this is. And what I really want is for you all to go out and feel it. Dario was right. When you start using Salesforce with an agentic interface, the amount of value that you get from Salesforce just explodes. It's an incredible experience.

Okay. So we've talked about AIforce. It's time to talk for a few minutes about Agentforce. Agentforce has been this incredible platform for us. I think I was the one launching it maybe 2 or 3 years ago. I've lost track of time, right up on this stage. And then in that time, we have over 30,000 customers on this platform. We're doing 7 billion AWUs. We've been adding more and more to it, like agent script and voice, which are GA now. But we've kind of learned something as we've started putting this out. This is built as a platform. So any of you can come and build your own agents.

But what we learned is that building agents, it's actually kind of tricky building with a probabilistic system is a new muscle, is a new skill that we all have to learn. And so we wanted to solve for that. So this year, we have been hard at work building out-of-the-box agents that sit on top of Agentforce that anyone in the organization can go and implement and deploy. Anyone, not just IT, but your business users as well. And that starts with agents like Hunter, Marc has talked about this a little bit, your outbound sales agent out there responding to your leads, hunting them down and trying to turn it into real pipeline or Piper, which you're going to see in just a moment, which sits on your website, responding to questions that all of your customers might have.

And then, of course, we have Kacey when you need to get help, you can go see Kacey right now and help salesforce.com, you can deploy this on your own website so that your customers can get help, but your customers aren't the only ones who sometimes need help. Sometimes your employees need a little bit of help as well, and we have a Paige for IT service and HR service. So if you're laptop breaks, you know exactly where it go. And as Marc pointed out, we're incredibly excited about our all-new customer agent, Fin, which is just an incredible agent. I hope you get down to the camp ground to see Fin. And now to bring Agentforce in all of these agents to life, I'm incredibly excited to her first Dreamforce main stage appearance. I think you're going to like her. Please welcome Katie O'Neil.

Thank you, Patrick. Siemens is a global leader in industrial AI. Their technology powers buildings and factories and transportation around the world. Over the last decade, Salesforce has really become an integral part of how they run their business. And today, they're driving success with Agentforce, a family of agents working together across [indiscernible] and we're going to show you these agents in action. So let's step into a demo.

We're going to start here on the Siemens website. At Siemens, they rely a global ecosystem of partners to bring their technology to their customers. But when one of those perspective partners arrives on the website, they could be left searching their pages, filling out static forms and ultimately waiting for someone to get back to them. That is potential revenue slipping through the cracks. The Piper changes all that. Piper is Siemens inbound pipeline generation agent. You can think about her like their newest employee on a single mission, converting inbound buyers into quality pipeline. Piper, she is always on. She is super easy to implement. Every customer goes live in 45 days or less, and she is working 24/7.

So let's jump in. Let's engage with Piper. Welcome back. How can I help you today? Piper, I got an e-mail from Siemens about becoming a distribution partner. Can you tell me a little bit more about the program? Absolutely. Based on what I know about your business, I think there could be a strong fit with Siemens' Industrial Automation distribution partner program. I pulled together a quick overview based on what's most relevant to your business. How incredible is that? Piper knows who I am because I click through an e-mail. She's grounded in the context of Siemens Salesforce. So she knows everything about my role and my company, combine that with all of her knowledge of Siemens messaging and positioning and Piper just spun up a super personalized presentation just for me in seconds on the website.

This is the dream. You have a hot buyer and Pipers ushering them along in their journey. Now this is peak bites, and I do want to take it one step further. Thanks, Piper, what would the commercial model look like for a distributor of our size? Great question about the commercial model. It's definitely worth exploring in more detail. Let's get you in touch with the Siemens representative who can walk you through it. Go ahead and pick a time that works for you. Feel free to pick a time that works best for you. Okay. So Piper knows that a question around commercial signals buyer intent. Combine that with my account context, Piper knows that right now is the time to bring in a seller. So she looked inside of Salesforce, saw that Ryan Smith is my line account executive. And in just a few clicks, I can book that time.

Boom, pipeline was just created. No forms, no waiting, hyper, just generated quality pipeline for the Siemens business, that was totally incredible. Now that's one area of the business where agents are going to work. Let's expand Siemens digital workforce. Let's talk about another agent, an entirely different area of the business, supply chain. For Siemens, something like supplier onboarding, it is a process that can take days, a single disruption. It can cause a huge domino effect that creates production delays and it can cost the company millions. But that's what we're solving with Marshall, our operations agent. Marshall is going to take that entire origination to and bring it down to just hours. And here's how. All we have to do is give Marshall the name of a supplier and he's going to go to work in the back end. We're here in Slack, but Agentforce meets you where work already happens. So if you work out of e-mail or Microsoft teams, Marshall will come find you in those channels.

And you're watching Marshall go to work. He's doing things like collecting information. He's reviewing documents. He's even going to pull in a person when something needs human review. Now here you're watching his getting though the merit. But if you'll notice, there's one area where he's getting stuck, and that's actually creating the supplier inside SAP. And in fact, this is where a majority of AI projects hit a wall, which is actually getting our agents to work inside of back-end systems. So today, we are so excited to announce new capabilities with Marshall that is going to do exactly that and let's show you. As Marshall jumps in these back in systems, the first thing he has to do is he has to learn. He has to learn this process supplier onboarding. It is not a trivial process. So we'll give them access to an SAP sandbox.

You can think about a sandbox like a classroom. It's a safe learning environment where Marshall can go through and safely learn this process without touching production. You're watching them do things like learn where to click, which fields are required, but he's not just memorizing those things. He's using LLM reasoning to really understand the business rules behind that process. So here, for example, Marshall will learn that a U.S. supplier actually needs an additional form before they can continue onboarding. Now this will process of learning, this is what would take a new employee weeks to learn. But Marshall, he just learned it in 90 minutes. I want to see do with all the learnings. Marshall will package them up, create a library of trusted actions, and that is where AI reasoning becomes deterministic execution.

What does that mean? It means when it comes time for something like supplier onboarding, Marshall is not just improvising. Marshall is executing against the same trusted actions every single time, so you know you can rely on him. Okay. So we just watched Marshall learn this entire process of supplier onboarding. Let's put them to the test, let's see them go to work. Now in this next part of the demo, I do want to turn our attention over to Gus or demo driver, Gus, give here in the audience a quick way for me, but actually Gus, waive both hand because look at this, Gus is actually completely hand-free in this demo. Gus has enjoyed his cup of coffee, well, Marshall goes to work. Look at this, this is an incredible. Marshall is in SAP. Marshall is updating every single field completely autonomously. This is incredible. This is where agents are at their best. Repeatable work that has to be done right every single time.

Now back over in Slack, all your employees are going to see is that this supplier has now just been created, has been onboarded and is activated. And as something inside SAP changes, no problem. Marshall will go back into that classroom. He's going to relearn and adapt, and that is trusted process automation. That is how Siemens is putting agents to work across their business. That's how they're using Marshall to take supplier onboarding from days down to just hours. That is how they're using Piper to generate quality pipeline. This is Agentforce. This is your digital workforce or here during first. We are so excited for you to get hands on. We want you to walk away with your own family of agents. So come find us in the learning labs.

And for now, Marc, I will pass it back over to you.

All right. [indiscernible] for Katie. Great job, Katie, well done. Great demo. Well done, fantastic, amazing. Katie came from our qualified acquisition. Qualified team, stand up and take a hand, qualified team. Piper, amazing, what you have done, built Piper, great job. Great, great job.

Okay. My friend has come from Munich, Germany. I was just with him. Welcome, Roland. Fantastic. Great. Nice to have somebody in my size here. People don't know my families from bot and bought in and Bavaria. So we have a lot of fun together. We've known each other for a long time. We're so grateful to have you as part of a Salesforce Ohana and your first Dreamforce, isn't that right? What's your initial experience at Dreamforce so far?

I know you just throw a great party. Great people. And Katie, thank you very much. Great job. And I'm so happy to learn that you share talent.

Great. Well, we're just starting. We're just starting. Now -- okay, Rohan, give us your vision because you understand these deterministic layers, but also these probabilistic layers, how AI works in the enterprise. You're building products yourself. Siemens is an incredible company. you built some of the best, most exciting technology across every industry, whether it's industrial or even health care or energy, it's amazing. I think all I said some time of our life or found ourselves in a Siemens CT scan or seeing how huge energy plant that you have built, it's amazing what your company does. Give us your vision of how you see this technology impacting Siemens.

Well, I mean, you said it. We are building the real stuff. Every third manufacturing line in the world is automated the Siemens technology. We have every other electron is touched by Siemens technology or call by it and so on. So in short, what we do is we combine the reel in the digital world in the age of AI. So we built a real thing. This is the hardware part. But we also build digital twins of the renew world, which behaves really like the real thing. So binding that is the huge potential and AI is accelerating. So we talk about industrial and you phrased it perfectly. This is bringing a probabilistic technology into a deterministic world. Hallucination does not really work on the shop floor, as you can imagine. So -- and that's the exciting part about the journey which we started. I mean, the Salesforce is the backbone for how we serve our customers in selling, in maintenance, service and the like.

And the onboarding of -- I mean, we have 120,000 suppliers, 65,000 partners in count. So that's huge. But we don't stop there. So you are the Salesforce of truth for customer, the enterprise AI. With our technology with team center, this is the backbone for products, the products in instructions for manufacturing for service, they're single source of truth of the product. Now we bring that together. So the enterprise AI meets industrial AI, and that helps really our sales guys, our service guys to serve our customers much, much faster and better. We -- basically, we give a virtual engineer next to each sales rep or service rep. And that's just the start. I mean, Dario said it, even if the models would not change from now on, I mean, they didn't see most of the industry data yet. And this is really where the real world kicks in makes a big difference.

Yes. That is a key point. I want to ask you to that point, you and I have talked for years and years because you helped us build this amazing children's hospital, which is a few blocks away from here. So one of the things we've always thought about is the CT scanner is going to be not just kind of -- it's going to be, I would say, more intelligence, more of a diagnostic device. It hasn't exactly happened, right? Radiologists are still in the loop. Even though Jeffrey Hinton said in 1995, we're not going to have any more radiologists. We have more radiologists today than ever. Tell us how is AI and that world coming together, I know you thought about this exhaustively.

So the thing about AI is you come from a world where you just have a copilot somehow helping you in really embedding AI technology deeply in your workflow. You have to rethink the way how you design products, how you produce them and how you operate them deeply. It changes user interface. I mean, you said it perfectly. It changes everything and the whole process behind. The big thing in that particular case is data fusion because currently -- I mean, it's -- we're already there to analyze CTM scanners with an AI agent to really help the doctor -- the doctors, obviously, finally in the loop. But if you fuse data from the blood diagnostics from your scans, even the medication you get, this makes the way how you treat people or how you actually act before a disease strikes, bringing them completely to the next level. By the way, eating habits does its own.

I want to ask you one critical question is this, Roland, when you think about your vision for your company or for all your companies, you have such a portfolio of companies now, over the next year, call it, what's the biggest change do you think that your customers will begin to see?

They will see that new superpower, which is a channel purpose technology will change everything what we do, the way we design products, the way how we manufacture and the way how we use them. And that happens, Dario, you have said it, in an unprecedented speed.

It's great to have you at Dreamforce. Dreamforce, welcome. Glad to have you. Please welcome him. Thank you so much. Great job. Okay, MaryAnn, take it away.

We are all here at Dreamforce to answer one big question. Will AI actually work for my business? You know what is at stake. A lost customer, a missed quarter, your reputation. Generic AI is basically like an amazing race car with absolutely no wheels. But the Customer 360, well, that is the wheels and the steering wheel. Why? Because of all of you, you have for the last 27 years, put all of your business logic, your data, your processes right inside of Salesforce. And the Customer 360 brings all of that alive. And then what we did, we took that amazing foundation that you built, and we built on top of it. Our product team has been working so incredibly hard innovating every single cloud. We've completely reenergized every single business, and we put agents and AI at the heart of it.

Now I know that we have 1,600 sessions, but I have 9 super favorite features that I want to walk you through. So are you ready? Already, a lot of start with sales. What's the absolute best thing about sales out? Well, it is Hunter, of course. And sorry, Patrick, but the best part of Hunter is not just that it's all new, and it is an outbound sales agent. It's that it is built with your data vendors included. What does that mean? Yes. ZoomInfo, demand base, Apollo, your data vendors are built right in. That is amazing. This agent has everything. But I am now on Service Cloud and my family, well, we are kicking butt because CCOS is going live in October. Now this person is looking and be wondering why we're putting CCaaS -- sorry, Contact Center as a Service is important because it's your contact center.

It is your telephony. It is your CRM as one system. Every single person needs this. Why? Because you don't want to think about technology, you want to be focused on your customers. I think I hear angel singing, no more tool stitching ever again. And what about marketing? Well, marketing is crushing it and they're all new campaign agent, it actually builds and iterates campaigns by itself. Now how many marketers love manually building 500 campaigns in the middle of the night before train for us? Yes, no one. So I am sorry, the midnight snacks business is now officially closed. Marketing cloud is really taking us to the agentic future. And the new Commerce Cloud, well, it has a shopper agent, which, frankly, is way too good.

I had one conversation, and then I ended up with 10 outfits for this keynote. Because it looked at the shipping information, the inventory, the even maybe backups for what I was going to buy. So I said yes to the dress and then to the suit and then to another dress. Thank you, Commerce Cloud. And if anybody here has people who are fixing things, physical things out in the field. Well, you know that field service is where it's at. And the best thing about field service is voice to text. Now who loves taking off their gloves, using greasy fingers, touching a phone while you were doing the job. Well, voice to text, you have no hands that you need to use. You can do more paperwork without your hands, you just speak and the form literally fills itself. That is amazing.

Okay. Now hold up. We just launched IT and HR service last year here at Dreamforce. I happen to run this business, so I'm not allowed to pick favorite children. But if I was, my favorite child would be our brand-new CMDB. Why? Every single IT organization knows that the CMDB, it is the brains of your organization. It's how you map your asset. And ours will even update itself, you're welcome. And with Revenue Cloud. What's Revenue Cloud? Of course, it's quote to cash. And the all-new renewals agent, I absolutely love that it automatically figures out what accounts you need to go reach out to and it even does the upsell. That is amazing. Anyone found their favorite yet? Keep watching because now we have Tableau, Tableau gives you a semantic understanding of your entire business that's live for your teams and your agents.

And what is even better than that? Well, proactive intelligence is better than that. Who wants to know the story without even having to ask for it. Thank you, please. And then there's industries. Every single one of us, we run our business inside of an industry. There are regulations that maybe aren't just ours but our regulators. The industry's team is amazing. They have passed all of that into industries that know-how is built into Salesforce. You can do what you want and make sure it's compliant every single time. And I'm going to steal this race for Miguel, but we have an embarrassment of riches. There are over 500 agents and actions built right in every single industry, every subsector. So if you are wondering, dive right in today. You just saw the amazing reimagined Customer 360. It is on fire.

This is amazing. So as you can tell, I'm a little obsessed with product, but what I love more than product is I love the people I get to work it with. So Rohan, can you stand up Rohan just joined us. He is our new President of Platform and Engineering. This is his first Dreamforce, everyone. Can you tell us what you're working on? I've heard you have an amazing announcement for us.

Oh my God, it's amazing. It's -- for the last 3 months, I've traveled around the world, met a lot of customers. And one consistent piece of feedback that I've heard from everyone is that they want a choice for AI. They essentially want to pick the AI model, which is best in terms of price performance for the outcome that they are seeking. And given that feedback, I'm so excited, are you guys excited to introduce Cowork Salesforce's first CRM reasoning model for Customer 360. It's interesting. You just step back and think about 27 years of product innovation in CRM, deep understanding of the customer scenarios. and almost a decade plus of AI and data research, all that learning and all that value has gone into Cowork. Cowork is designed to support long-running agents that execute multiple tasks and complete complex outcomes. So think about a sales agent that's trying to figure out the next best step to take to actually pursue a complicated but a very valuable sales opportunity.

Cowork can help with that. Think about a service agent that's actually figuring out the next best step to take to handle a complicated customer case, a customer support case to get the customer to the right place. Cowork can help with that. As a part of designing Cowork, our commitment to customer trust has always been maintained. It's something that's extremely important to Salesforce. Cowork has always been trained just on synthetic data built and extending Nemotron from NVIDIA, another phenomenal company, right? And the reason that's important is because our commitment to customer trust has been carried forward. Not a single bite of customer data was used. We can't get to see what you build with it. and to talk about where we are going with this incredible partnership, let me hand it over to Marc.

All right. Great job. Please welcome Rohan Kumar. We're so excited to have you, Rohan as our new President of Engineering and Platform, but we are especially -- please welcome Jensen Huang, the CEO of NVIDIA is here. Great to have you, Jensen. Thank you for being here.

We all have to learn how to walk and talk at the same time. How are you guys doing this?

Jensen, we can do that. Yes, let's walk here come up. Let's walk and talk. I do the training myself, Jensen -- exciting thing. You have to take a left now. So Jensen, congratulations on everything that you are doing for the whole industry, the GPU, it's a huge success. It's transformed everything. You're sitting next to David Kirk on our Board, who I think was at NVIDIA for a while as well.

That's right. Love David Kirk.

And now give us your vision of the future, where are we going? Because of course, you've got these critical components. You've got the GPU. And now you've got the #1 open source platform in the world that companies like ours can fine-tune. How do you think this can go ahead? Really we can go this trader do this.

Well, you had to stop so you could think. I could do this all day. Yes. And so first of all, as you know, we're going through a new industrial revolution just as a few hundred years ago -- WI think electricity, we could power everything. With the Internet, we can find -- with Internet, you can find anything. And now with artificial intelligence as an infrastructure layer across the planet, we can know everything and do anything. And so this layer, this layer, you want it this way -- now you got it -- this is great. This is good. Yes. And so now that's the first part. Now the work that we do together, what we see is that every company, every enterprise, every country would become an AI company.

Everybody would be. You're an AI company. I'm an AI company. We're going to have -- it's going to be Agentic Enterprise. And so what we did was we said, listen, we took the GPU, and we evolved it into these giant systems that are now essentially full stack AI factories. And so -- just follow me. We got this. I got this. And so listen, so what happens is we evolved our company from a GPU company to a full stack AI infrastructure company. We build out these giant systems that are essentially AI factories that turn electricity that turns the data that all -- that are stored in all the enterprises and we turn it into intelligence. And I was just amazed that the way you were doing it. Was kind of -- did you read that? I don't think so. You guys were just doing it. So anyways, so what we did was we created this full stack AI factory. And then on top of it, we said listen, how can we turn every single enterprise into an agentic enterprise?

And so we did several things. The first thing that we did was we created the frontier AI models. Nemo transfer language, but we also lead 5 other frontiers, Nemotron. Now you guys are going to make me nervous. And so we have Nemotron -- we're going to cover everybody. Is that a metaphor for the AI industry, do you think -- we can go anywhere to go. And so the first thing is Nemotron is the language model, Kosmos world foundation models at the frontier, it is the frontier, Alpamayo for navigation systems, autonomous vehicles, for example. We have BioNeMo with a protein complex at the frontier of protein models incredible and then, of course, GR00T for human robotics models. And so we're at the frontier of these models all completely open. That's the first thing that we do. The second thing we did was create agentic harness system, and it's completely open. You could put your model inside of great, amazing harnesses around it and the run times that do it.

And then the third thing super important, is this thing called Open Shell. It is the world's first secure sandbox, but also secure run time. And so that you could deploy these agents at scale across enterprises safely. And so these are the things that we did. And now we can turn every single enterprise, every single company into an AI company. And of course, we believe we see a future where the world uses closed models, but every company would also be able to build their own private models, your own customized models, which is the work that we're doing with you guys. Salesforce is going to be a company.

Jensen, we have been seeing an incredible debate for the last day on pacing and safety and the future of these models, you had some time to think about what everybody is saying and kind of bring it all down to us and explain how do you see that?

Well, safety is paramount. In a lot of ways, it's job one. However, safety is an engineering problem. We're developing software after all. We're developing computing systems after all. It's complicated computing system, but it's ultimately a computing system. And so the first thing is to make sure that we create the test environments for testing these complex systems. And those test environments have to be built in a good way, build in a safe way. And then the second thing is we have to test these products.

If we're not confident about the safety of the products, like all companies, like you and I, any -- all the companies here, if you build a product or a service, and you're not confident in its functionality, capability or safety, then don't release it. And so that's a very obvious thing to do. You pace yourself until you are confident you're releasing something that the market would appreciate. The market forces are already there. We don't need any new laws. We don't need new regulations.

We just need companies to decide that when it's -- run as fast as they can. I think innovation, speed and safe products are not -- it's a false choice. You could definitely have both at the same time. So run as fast as you can. But if you feel at any given point in time, the company is out of control or the product is not going to be safe, take a pause and make sure you get it right.

What's your vision for the next year in the short term? What did you like to see everyone in this realm? We have 10,000 people, you can see all around with 10,000 people in this room. We have 10 million people online, 50,000, 53,000 people here at Dreamforce with us this year. I think this is your second Dreamforce. What do you want to see us all kind of get done this year?

Well, I'll say this first, and then I'll say last. And I'm pretty sure what we want to do is Agentforce every company. Do you guys agree with that? All right. So I'm a sales guy. The consummate sales guys, sales guy. And so of course, what I mean by that, of course, is that we really want to deploy AI. In order for all of our -- we were -- the people here, you represent a lot of different companies, a lot of different countries. You don't want to be left behind. This is too important of a technology revolution. It is too extraordinary of capabilities, engage the technology, learn about it. If it doesn't work for you right away, don't give up on it.

And the reason for that is because there's a massive industry working behind it, making it better all the time. If you just tested it today and you felt that it did 80% of what you expected, test it again in a couple of weeks. It's going to get much, much better. And so don't give up on the technology. The most important thing is don't get left behind. I think that the big picture though, is that in the last 6 months, there's no question AI went from the labs into what I would consider extremely useful AI you're experiencing, and I'm experiencing across all of our companies.

The surprising thing for a lot of people and wasn't surprising to me, okay? And in fact, you heard me say it, I was the first CEO to come out not in the software industry to say, the end of software is nonsense. This is going to be a layer on banker software -- this was -- it was a layer on top of software. This new layer on top of software is going to be agentic, it's going to make you use the software so much better. And so that's the first idea. The second idea is AI is going to destroy jobs. That's also completely nonsense. And the reason that is very, very simple.

There was -- you and I are -- we look young, the fact that we could -- we do by fact -- yes, the fact that we can -- we really do -- you're at that we could sit here and pay and talk at 100 miles an hour and keep them with the young people like that, they -- we are -- it is true -- look, I'm very good friends with both of you guys. The thing that it was managed by is, finally, look, when I'm with the 2 of you, I always feel like I need to stand on a chair. And literally -- I just wanted you guys know that when I'm on an airplane, I'm a lot more comfortable than you guys. And so through human evolution, that big is unnecessary. There are several -- it's about 1 million years of evolution between you and me. And so...

This keynote is making a dramatic turn.

Well, we're just -- we're going to connect it to today's entertainment. And from here, we're going to pivot right into comedy. And so Marc, the thing that's happening now is, of course, AI is becoming productive and useful. And as a result, the number of tokens that are being generated as skyrocketing. One of the -- some of the most important things, of course, is that closed models are doing incredibly well. In the last year, I think the number of tokens increased skyrocketed by 25x. And that -- a lot of that in the last 6 months.

What's amazing is that the open models in that same period of time increased -- out of that -- excuse me, the open models went from 30% at the beginning of that year to now some 70%. And so what that basically says is that people are both adopting closed models at exponential rates but also people are building their own custom AIs because every single software company is an AI company. And every single enterprise is going to be an agentic and AI enterprise. And so I think that these transitions are super powerful for the industry. As a result, companies are going to become more ambitious than ever. You're more ambition than ever. I'm more ambitious than ever. As a result of our ambition and with the productivity both we get from AI, the sky is the limit for us. The sky is the limit for our company. The sky is limit for every industry, for every single country. So engage in, I don't get left behind.

Yes. And I'll tell you, the sky is especially the limit for our Trailblazers. I think this technology, the way it empowers people, the way it takes people who can really make a huge dent in companies and transfer them. The technology can partner with them to do this in an incredible new ways. That is what I'm really excited about -- we could not be more grateful to us for coming to Dreamforce being -- being a part of this. Please thank Jensen Huang. Thank you, Jensen.

All right. Amazing. How about that? Wasn't that fantastic? And that -- with that, I want to bring back our good friend, Rohan Kumar. Rohan, let's hear it.

Thank you, Marc. That is amazing. Well, this is not going to be a comedy session. But I want to talk about something that's foundational to everything that you've seen today, data and enterprise context. Enterprise context is anything that makes you ones uniquely your business. These models are very intelligent. They're extremely smart. They know a lot about the world. They know nothing about your business. They don't know your customers. They don't know your products. They don't know your employees. They don't know your transactions, your conversations. They don't know your business workflows. They don't have the history and the context of how your business functions.

And guess what? At Salesforce, we have an incredible suite of products that helps you create this enterprise context, manage it, secure it and make it available to all the agents across your enterprise. Let's get started. The first step is to get your data ready for AI. Before your agents understand your business, you need to understand your data. It's actually extremely complex. Data is locked into silos, data lakes, warehouses, storage systems, applications. Informatica is the product suite. It's the world-class product suite that helps data leaders discover their data from all these silos, clean it, curate it and make it ready for AI. So that's the first step us Informatica to get your data ready for AI.

Now once your data is ready for AI, you basically need to sort of create the trusted context. And the way you create trusted context is today to 360. So what is trusted context? A customer of yours [indiscernible] support case, that is data. deeply understanding that customer, what is their business? What are they trying to accomplish in the moment? And what can your agent do to take the next best step and help them, that is enterprise context. And with Data 360, you can use the 0 copy capability to get the data that's outside of Salesforce. We know there's a lot of your enterprise data that's outside of Salesforce. And you can get all that data, synthesize the enterprise context and knowledge without actually physically copying that those data assets. So that's how you create the enterprise context.

Now once you have your enterprise context created, you need to create the business semantics. Every company has a unique way in which they think about the business conversations, what does revenue mean? What does ARR mean? What does customer churn mean? What does customer health mean? These are not just some static fields inside a database. These are business metrics, business relationships and logic that have been built over years. And Tableau is the most amazing tool for you to go create those business semantics, augment your enterprise context and make it available to all the agents across your enterprise. Now you've got your data AI ready, you essentially created the enterprise context and you got the business semantics.

These agents that have access to all this information are extremely powerful, but put yourself in the shoes of the security leader. They create an amazing amount of security risk. What do these agents do? Like what are they accessing? Can they leak my confidential data, my enterprise context. These are serious questions. That's why I'm so excited to introduce Salesforce Guardian. Salesforce Guardian is the security suite of products. That's the evolution of Shield and trusted services that focuses on 2 things. Agent identity, how do you discover agents that are going rogue and data security. So think about data classification and data life cycle protection. Those are the 2 things that we are focusing on or solving with Salesforce Guardian. And of course, our commitment to zero data retention. You saw Marc speak about it multiple times.

It's very important to us that even when models reason over a confidential enterprise context, that's your IP, we will always protect it. All right. That's how you secure your AI. The final step, essentially, just think about the IT leader. Now you have a digital workforce that's coming together, hundreds of agents, thousands of patients, who knows, millions of agents across all your teams, platforms and enterprises, how are you going to manage them? How are you going to discover them? How are you going to control them, especially cost control. That's where MuleSoft has been upleveled into Agent Fabric. IT leaders have trusted MuleSoft for decades to manage their APIs that connect their business. With Agent Fabric, you can manage all your agents using 1 single industry.

All right. So you've seen Informatica gets your data ready for AI. Data 360 builds out your enterprise context. Tableau builds out our business semantics, Salesforce Guardian help you secure your AI and then MuleSoft and Agent Fabric helps you manage and govern your AI. Let's see how this comes together at a customer, Adecco. Adecco is an agentic enterprise that actually is helping candidates who are looking for jobs, mapping them with their active opportunities. Think about it, millions of candidates every year looking for new jobs, right? And to get this right, the Adecco agents have to rely on trusted context that's built on Data 360 and Customer 360. It's incredibly exciting. Just this year alone, based on the scale, they manage 2.6 million agent interactions with these candidates.

I'm going to show you how Adecco is actually building these agents in a very trusted and scalable manner. But before I do that, I'm going to pass it to [ Mara ] at least show us the agent experience.

Awesome, Rohan. Let's do it. Who's ready? Okay. Okay. So Adecco is this huge staffing company, as Rohan said, and they connect people all over the world with their dream job. But the scale of Adecco, it's massive. They have over 5 million candidate applications every single month. So the scale is great, right, but it's too much work for the recruiting team to handle. What did the Adecco team do? Well, they went ahead and they brought on a recruiting agent to manage all of these candidates to pull them through the funnel. And they named this agent Oda. They wanted to personify this agent, so it felt like an extension of their brand. And Oda is built on Agentforce voice. So she can have live phone conversations with candidates all around the world 24/7.

So let's see it in action, who's ready? Let's do it. Okay. So we're going to do a little role play right now. Let's pretend that I just replied or applied to an IT role. Wish me luck. It's not my forte. But my application, it's stuck, it's stuck in the process. So I want to give Oda call, and I want to see what's going on. Hi, Mara, I'm Oda, Adecco's AI recruiting assistant. I see your IT support specialist application JR-325421 is paused due to an expired CompTIA certification. Have you renewed it? I have. My confirmation number is MP-2890. Am I able to move forward in the process? Let me take a look. I was able to verify your new certification. I'm updating your application now. Is there anything else I can help you with? No. I'm all set.

Okay. So what did happen? I just had a live phone call with this recruiting agent, and it was a really productive experience. I was able to get my problem solved. I was able to move through the process. And the question is, how? How did that work? Well, it's because of all of the rich data in Data 360. This is your agent single source of truth. Marc said it earlier, this is the data that drives your AI. So Oda, she was able to see everything about me. She saw what my name was. She knew my identity, she saw my job application. She knew which role I was going after. And she saw that certification flag. She knew that I needed to renew it and complete that in order to move forward in the process. So this is a really great candidate experience for me, right? I'm feeling good.

But the power of agents, it's scale. That personalized experience that Oda just gave me as a candidate, she could give it to hundreds of thousands of candidates, all at the very same time. So speaking of scale, what's really important is that the business leaders at Adecco, they can see the impact that she's having on the business. It's not just about one candidate interaction. It's about knowing that she's moving the business forward. So here we are in Slackforce. And this is a live interface that's pulling all of the information of what Oda is doing with their candidates. I look at this, and I'm like, wow, auto's going to work, right? At has been busy. I can see that she's having conversations all around the globe, all the way from the United States to Spain to France. And more importantly, I can see that she's moving these candidates through the funnel.

Because when you think about agents, it's not just about knowing they're having all these interactions, it's about making sure that they're moving your business forward. So I can see she's pulling these people through to hire and she's filling these jobs. So I just showed you 1 example of an agent, right? I showed you a recruiter agent that Adecco has used to engage these candidates at scale. And companies like Adecco, they are deploying agents all across their business. And all of you guys, right? But the next major question is how should I think about managing these agents? And how should I do it with trust front and center? So Rohan, I know this is top of mind for you, break it down for us.

Thank you, Maura. That was wonderful. So now you want to pay attention. I know it's towards the end of the keynote, but you want to pay attention at this time because I'm going to show you 2 new product experiences that have never been showed at Dreamforce. Agent fabric for the IT leaders to manage the agents and then Salesforce Guardian for the security leaders to secure their agents. So what you see in the screen over here essentially is the homepage for an agent fabric, right? It essentially has scanners that are actively scanning your enterprise continuously to discover new agents as they come up and then agents are becoming active. Let's go take a look at the agent staff.

Here, you see the list of agents, all the agents that are there in the enterprise, and I want you to focus on the provider column. You see their agents from Azure, Microsoft, AWS, Google and Agentforce, of course. So every provider's agents are actually managed by agent fabric. Now let's go a little deeper into the other recruiting agent that Maura spoke to. Here, as an admin, I get a very clear view of all the skills that the agent is using, how to candidates get qualified. Does the agent look at the verification status. They have listed a whole bunch of things on the resume verifying those are actually true. Let's go a little deeper into the lineage. Where is this agent getting its grounding data from. Remember, enterprise context is how the agents become deterministic. This essentially shows you how. And the beauty is we have deeply integrated Agent Fabric with Informatica. So all the catalog information in Informatica, which -- where your catalog was discovered across your enterprise, those sources show up, and so the IT admins can pick what the agents get connected to.

So that's the first issue. The next thing, as an admin, I'm worried about is, hey, how are these agents performing? What are they? Like what does the performance look like? And with agent fabric, you basically see a very clear view of the average latency. So every request, how long does that take to complete? What's the rate of these interactions, agent interactions. That's important. Policy violations, like is the agent going rogan should something be done about it, and then the total number of requests. This is important from a cost management standpoint. It's great that the number of requests are increasing because you want the business to grow, but how does that impact cost? Let's go take a look at that.

So in the cost management in Agent Fabric, it makes it really easy. So you have these things called Agent Fabric wallets, which essentially is a way for you to budget your spend across multiple agents. So here, you can see [indiscernible] sort of hitting words the end of the budget for the month. And Agent Fabric essentially gives me a way to optimize my agents. It's observing your agent interactions, what your agent is doing and using AI, it's figuring out certain optimizations that will help you reduce your token costs. So it's pretty amazing. So I can just go click, view the optimizations for data, understand the regulations being suggested if they have any behavior changes in okay with that. All I need to do is just go apply the change, and that's how I manage my cost. So there you go, Agent Fabric essentially helps you manage, govern and control your costs.

Now let's put ourselves in the shoes of the security leader, right? Like I said, the big thing that agents change are agent identity, really trying to understand the risk associated with the agents and data security. And those are the 2 big areas that Guardian is focused on in terms of new innovation, right? We show you all the agents that are considered risky based on the actions that they're taking. Then across your data assets, which data assets are not classified not protected, so the security leaders can take action. Wow. So that covers was Guardian. We've shown you a lot within Data 360. And guess what? Across team force, there are many more sessions that you can go watch to learn go much more deeper. I have a keynote that's coming up where I'm going to talk about the evolution of Data 360 to what we call the enterprise AI harness. This is going to be an amazing future for us.

And with that, let me hand it over to Marc.

All right. Great job, Rohan. Great job. Rohan, really, really great. And Denis, please welcome Denis, the CEO Adecco is here. Great to have you, Denis, [indiscernible] of me. We're really excited. We had our friend from Germany. Now we have a friend from France. So Denis, just give us the vision for how you are using Agentforce interviewing all these candidates and automating your companies and building new companies like our potential? Give it to us.

Well, first of all, thanks for having me. And it was great to hear Jensen saying that jobs are going to remain in the world, right? That the massive destruction of job is not going to happen. So actually, what happened is we started to identify the business by thinking about what was happening in the company. We are a people company. And recruitment is about people talking to people. And actually, our recruiters, we are doing a lot of things but not enough time spent in the human to human connection.

So what we did, we didn't start where it was easy. We started on what mattered most with the core process. We started on the core. But on top of that, we had a growth mindset. We wanted AI to be a growth engine for the business by generating a higher level of conversations with our clients and our candidates and with Agentforce, it's now possible 24/7. That was a life-changing moment for us. On top of that, the time spent -- the time saved. Now our recruiters saved 35%, 40% of their time, thanks to a TKI. And what they used that time to create much more qualitative conversations with clients. And when you do that and with candidates. And when do you do that, you grow the business. We have placed 20,000 more people year-to-date, thanks to better conversations. And in doing that, we also create joy at work because people are focused on what they love most.

And the last thing, because you mentioned our potential. I think when we started to work together, we had a common vision that AI had to happen with people and not to people. And that common vision created r.Potentia. This AI native platform that is there to guide our clients. Many clients still struggle about how they can scale and struggle on how to decide what work goes to AI and what work remains done by humans. And our potential is about that. It's about core business to Inventure measuring outcomes, this is fundamental and also bringing people along. So -- at the end of the day, if AI wants to be successful and if we want to measure and bring the returns, people cannot be an afterthought. We have a common responsibility to create a human-centric AI. That's what we are about.

Denis, great job. Thank you so much for being here at Dreamforce. More grateful to Adecco and you and for your leadership and everything you're doing. Thank you very much. Well, this is the moment to start your Agentic Enterprise journey. We have just an incredible show set up for you. You're going to have an amazing opportunity to get going. Dreamforce is where everybody transforms and we are so grateful to each and every one of you for joining us today at Dreamforce and spending the next several days with you. Welcome to Dreamforce.

Good afternoon, and welcome to the Salesforce session at Goldman Sachs Communacopia, I'm Gabriela Borges, delighted to have on stage with me, Bill Patterson, President and Chief Commercial Officer; Rob Seaman, EVP and [ GM of Black ]. Gentlemen, thank you for your time.

Thank you for having us today.

Bill, I want to start out with a little bit of some mic-drop movement from 2 weeks ago now with [ Dario and Mark ] on CNBC talking about the launch of [indiscernible] and there reminds me a little bit when we were chatting with the management team earlier this year, they were telling us about the anthropic Sasser video, where athropic talks about the cornerstone of their front office stack being sales force and almost the different pieces that they built on top of it. Tell us a little bit about how that relationship came together. And it sounds like you've essentially commercialized what would have taken a fair number of engineering resources to get off the ground. So would love to hear more.

Yes. I think -- thank you for recognizing that was 2 weeks ago, which seems like a decade ago now. But I think what's beautiful about the world of kind of cloud and the world of Salesforce is our customers were not sort of pulling us together to work together more effectively. And Salesforce is a huge entropic customer, and Entropic is a huge Salesforce customer. So there's just natural synergies about these 2 technologies sort of working together in unison to drive ultimately better quality agents, better productivity for every employee and sort of just better synergy on the product side.

So it's always beautiful when -- because like we always say it like the #1 AI platform and the #1 CRM platform can come together because we can actually really validate a lot of like the core use cases that we serve in the common roles with one another in that sense. So I think it wasn't -- it was more than just an announcement. It was a product that was already being built, already being utilized, already being activated, already being deployed in almost every customer that we sort of see today, and it was just a natural fit for these 2 product organizations to sort of now work together and deliver product, which we now call Cloud force. So I think this is just a classic example of kind of 2 magical things coming together and delivering great impact for our users, and we're very, very excited about the early days there.

Is there anything that you've learned in the last 2 weeks since making the announcement that was unexpected or that you haven't predicted?

Not unexpected, but I think that the market and sort of the customer community probably breaks down into 3 tranches. One, there's a lot of just heavy early adopters that we're really doing a lot of innovation and building a lot of these connections and integrations already themselves that validated the pattern for why we wanted to create a product in this area. And so I think as part of that early adopter community, you've got a lot of great feedback, a lot of validation, a lot of sort of conviction about the magic of these sort of 2 products becoming a strong partnership with one another.

I think as you sort of fast forward to the second tranche or second community of users, there's still those users today that are like trying to figure out how to make this work and how to make it scale and how to make it economically viable for organizations of their size. Not every small business has an endless token budget, if you will. And so really making sure that the economics work to drive the productivity and the performance of this offering is probably that second community of customer that now needs a little bit more knowledge and know-how to make it work and make it work right.

And I think there's a third community that has a lot of skepticism around sort of what does AI mean for them. And as you travel around the world in different communities, different geographies, not here in San Francisco, per se, but as you go to like other countries around the world, there is a little bit more AI hesitancy. And so that is a final community that does need to be convinced that this is not a destructive force it's actually forced to fundamentally transform work. So I think as you take a long view around what the Cloud First opportunity means, the early adopters are going to deploy it fast, those that are sort of trying to figure out the economics and the true math of how this works at scale, I think, are still needing some convincing.

And then that last tranche is narrow that we have more responsibility as leaders to show them that this is a strong sort of solution for their future and not, like I said, that cannibalistic force for eroding work as we know it.

I want to ask you about 2 of the bare cases that we've heard in the last 2 weeks. The first is the Fox in the henhouse bar case, which is, look, tech partnerships come and go, they break down all the time. and essentially through a quarter for us, Entropic is now getting a foothold in the front office tax of all sales forces installed base. How do you think about the risk that -- however many years from now, anthropic turns around and says, thank you, Salesforce. We're happy to take it from here.

Yes. Look, I think that the fundamental mission of both organizations is different from 1 another. The mission of Anthropic is really to prove that kind of AI can have a really ubiquitous sort of kind of moment in our lives in our businesses. And the role of sort of Salesforce is fundamentally to kind of make businesses better businesses. So we have some relationship in that sense that to -- of what they innovate on and what we innovate on. They're very, very different technologies, very different pace of sort of innovation and also better sort of just different responsibilities of what we need customers to sort of know, learn how to do and manage to be successful with these offerings.

There's been all kinds of new innovation that has always been the death of a CRM system at some point in time. The first one was like a website. Oh my god, that's going to eliminate how we do kind of marketing and marketing outreach. Now it didn't transform I didn't do away with CRM. It just transform CRM to be a world of outside in engagement, just as much as it's been about inside out sort of operational responsibility. So I think it's -- there's a long list of companies and long list of technologies that we're trying to kill the CRM business. We're still here. We're doing well.

The second bear case, and we can bring Robin to the conversation naturally with this one, is the idea that Salesforce is essentially giving up the user interface? So the end customer now interacts with clouds. There are quarters port of call. And sales force essentially becomes such database argument from 18 months ago now. How would you respond to that, particularly UI bar case?

I go about it from a perspective. One, I don't agree with it. I'd start with that. So I think minimizing sales force or CRM in general, to a database I think is somewhat glib and a bit of an oversight into what's actually happened in the CRM systems. Like so much of what people have done is expressed how the company operates in the form of metadata in Salesforce and their CRM. And that, candidly, isn't very replicable and that's what people are excited about accessing through other services like Cloud force or Slack or whatever it might be.

Now on the argument about giving up the user interface, we -- while you may be giving up a bit on the amount of time people spend in the interface, you're actually increasing the amount of work that's happening through the services that support the interfaces -- and to me, that's the bigger bet is that ultimately, more work gets done through our CRM system. Like if you look at our own internal data on our usage of Salesforce, our usage of sales force is skyrocket within Salesforce. And but the overwhelming majority of it is happening through Slack and through cloud. And so the amount of work that's actually happening and the quality of work that's happening has dramatically skyrocketed, and our productivity has improved, as you've seen through our financial metrics. But it's all because we've moved to this headless manner in a way, I think the tilt umbrella makes it a lot more convenient for users to get the work done.

Tell us a little bit more. You've talked about Slack as the operating system for work. What does that look like? What does it mean for your R&D road map to now be investing in Slack, not just as a communications tool, but as this essential gateway to all of the underlying system of record.

Yes. So when we think -- so the biggest asset we have at Slack is the aggregated attention of knowledge workers. So the average Slack user actually has the Slack app opened 10 hours a day and actively uses it 2 hours a day. So if it were a consumer app, it would be a great place to put ads, but we will never put ads in black. So we do not expect that to come from us. So it becomes an amazing place to distribute your software. So B2B software is increasingly being distributed through Slack as agents are. Now when you increase the amount of stuff that are coming at humans, humans can become distracted, and that defeats the purpose of that attention aggregation -- and so what we think with the role that we can play, if you look -- if you go back to original operating systems and what they did, they basically abstracted or hit the complexity of the hardware from the end user and typically how to use their interface paradigm about it.

That's effectively what Slack does for a litany of apps and agents that are now coming at people in addition to the humans that are reaching out to them. So from an operating system perspective, we think we can basically provide that overall design guideline and system that apps and agents can adhere to, we can abstract the complexity of the fact that there are thousands of them now competing for people's attention and make it a single place that people can kind of orchestrate their day and be a cockpit of, and so I think we ultimately become where the AI stack starts and stops.

You gave an example there of where it's been successful within the own sales force organization. And I imagine you have customers that you would classify as top quartile in terms of early adopters of Slack functionality. My question for you is what is the alternative? So the customers that maybe we separate out the cohort that still hasn't. So the customers that are leaning into AI that are not using Slack as one of their primary user interfaces. What does that look like? What's alternative?

I mean the alternative is alternatives from competing products that I think are functionally deficient. And so I think that -- the key thing about Slack, and I think you see this play out in all the major AI companies that use Slack, and I apologize in advance if I'm saying something that whoever -- if I'm saying something that I'm not supposed. But if you look at anthropic, you look at opening eye, you look at Versal, all of these frontier labs and AI-first companies operate their stock in a very unique way, which is they operate totally in public. I was talking to Boris from Entropic yesterday was so fascinating.

Every single anthropic employee has the notebook channel. That's basically a channel with them themselves, where they just think out loud through the course of the day. They knew this. OpenAI does this or sell does this. We do it at Slack. And what that does is that radically flattens an organization and immediately disseminates all information from that person to every other person in the company, but importantly, to all the other agents, and that little in Slack and the way that the software actually encourages people to participate in a very social way in a very public way is not something you see happens and the alternatives that are out there today full stop. I would challenge anybody, if we're going to talk to a customer that's using an alternative product, and we ask them how they do that. And this start like you have 10 heads...

Well, you got to stay on this. This is a really interesting rabbit hole. So what is Entropic doing with Slack that allows them to power this army of agents behind manifest? Tell us a little bit more about...

It do as much as seemingly possible in public. So every single product channel, feature channel is public for the entire company access and all the agents do access. Every single person then has what we call a notebook or brain channel. Versal does a similar thing if you talk to G from Versal [indiscernible] channel and they just think out loud in these brand channels. And so everybody in the company can see what GE is thinking. And every agent that's in your slack can see what Geis thinking because it's public. And I think increasingly, if you look at what -- Toby look -- I don't know if I'm pronouncing that right, the top if. He wrote a piece on Twitter called learning on the shop floor. And it was about using their river agent in flag, but only doing so in public.

And what's interesting is they've deployed their coding agent river, which is built on top of another harness, but they've deployed it in Slack and only allow it to be used in public channels and flag. And so what you're doing there is you're effectively immediately disseminating to the entire company, every single thing that's being built. So everybody can deserve like somebody else working on this yet. So like you radically duplicate -- reduce duplication of effort. But then you also learn from each other. Like the bare case to me on these individual contributor, AI tools, is that -- and Toby says this as well is that the AI learns and the person using it learns that their peer doesn't learn and then the company doesn't learn from it, right? When you move that into public like in a channel and Slack and then everybody else can learn from it. and it immediately becomes a dollar artifact for employees, but also for other agents in your slack.

So you've already given us one of the answers to this question on why Slack is the natural place for that activity to happen, which is it's already open 10 hours a day, it's active 2 hours day. What else is in the IP of slack that you think allows that use case to flourish?

I mean suddenly comes down to how we work. People are -- so One of the things we've traditionally said about Slack behind the scenes is our enduring competitive advantage is going to be the quality of our user experience and the customer love that we generate from that user experience. And the fact of the matter is like we have really high NPS, customers love to use their stuff. They tell their friends about it. It's viral, it takes off within a market. And so when people are asking their company for access to software, they're like, "Hey, can you put this thing in Slack. So we've been able to develop this kind of groundswell and grassroot support and pull from our users. I think that is really, really helping us right now.

You saw it play out a large social networking company last week, right?

This opens up an interesting trade here on monetization, and I'll ask it in a couple of different ways. So first with Slack. Historically, it's been a seat count model?

Either [indiscernible].

[indiscernible] And now you have any given user getting boat loads more value, especially for the AI pilot examples that you gave there. How does the pricing model change for the Shopify users of the world?

So it's interesting. So we have -- and Bill, obviously, feel free to chime in here. We have a series of plans. So we've -- I think we've been one of the more successful premium products that we still have overwhelming majority of what we call our invoice business, which is our sales business, our enterprise sales business, initiates what we call self-serve. So you started as a free team. You grow up through our plans through nudges in the product use more features of the product, start paying for it and eventually convert over into what we call an invoice enterprise plan that our salespeople help do with.

From a monetization perspective, I think there's a couple of things that we can do here is to one, like overall, like the total TAM for knowledge workers is around $1 billion at this point. So I think we still have a ton of upside. As it relates to user acquisition. But increasingly, I think you're going to see more agents in slack than you see humans, and that creates a fascinating monetization opportunity that we haven't totally tracked not yet on but I think it creates incredible upside for Salesforce. And so you can think of agents potentially as users. You can think of consumption-based mechanisms for monetization as agents are actually the access patterns for information and is locked by agents is very different than that from humans. And so we're in the early stages of that, but I could not be more excited about the upside for the [indiscernible].

And I'll build on that, Rob. I think as sort of Rob mentioned, what started as largely per user per month is now expanding into the agents that are sort of exploding the workforce. The usage that's exploding like never before. And even into the future, you're going to start to see kind of more things like outcomes that like organizations are sort of ruminating on or delivering with slack that sort of cortex of that work, that really become kind of opportunities for participating in more value exchange between us and the customers that we serve.

So as you think about sort of the monetization opportunities, they're highly elastic and they're highly sort of variable based on sort of what companies do with the software. And what is really fascinating about Slack is how much people do with the software and how we all embrace it in sort of our core work and workflow, and it just becomes almost like the place where I prefer to do my work as opposed to all these other systems that might have had to interface to do my job with prior. So I think the monetization sort of opportunity, what started sort of very small around the user base world now expands into lots of different sort of upside potential for many of our businesses [indiscernible] included.

And that, I think, goes back to something you asked earlier that I didn't perfectly answer when you were talking about our differentiation or how we might maintain this. I think that's our moat, honestly, is like the distribution that we've had, the compounding context flywheel that sits in Slack and just the more and more you use it, the more and more valuable it becomes -- and I think that's ultimately the moat.

Bill, I want to ask you a couple of broader monetization questions. The data point we got from Miguel on the earnings call was 5% roughly of the Salesforce installed base is on premium SKUs. And premium SKU can sometimes be a 60% to 80% uplift. Could you just break this down for us a little bit if we're a Salesforce customer, and we're not one of the ones that have already invested engineering resources and to making cloud-first happen without quote for [indiscernible] we say, Core4 sounds amazing. How soon can we onboard. What -- explain the 60% to 80% a little bit more, what is that uplift?

Yes. Let me give you -- it's sort of in the case of our traditional 2 core cloud Sales Cloud and Service Cloud, but -- when you think about all the software that's required to make a salesperson sort of productive, not only is oftentimes your CRM system for managing your sales pipeline, your context, your accounts, et cetera. But that often sort of expands into your sales performance management system, where my quota, where my commission, where my compensation sort of management systems are aligned maybe my sales and admin software, which is sort of help me kind of become enabled or effective as a seller. Maybe that's where all my content exists. I can come out and deliver pitches or proposal to my customers.

And there's also sort of all these extremely like sometimes it's called recording software or coaching software that helps me sort of in moments of working with customers. Well, the average sort of spend on a customer with all of those other technologies in the United States, it's about $1,300 for an average sales professional. And what we've been able to do is put a lot of this value into one price edition called our agent force 1 addition or -- we may have a new name for that next week, but I'll call it [indiscernible] today. That [indiscernible] addition is retails for USD 550. So the goal here is to really kind of take money out of the extremities for we know every sales organization has to recruit higher activate and make salespeople productive and performing and now put it into sort of 1 offering that's there.

And it's early days for us. We only started really thinking about premium levels of monetization really the last couple of years. And so as Miguel was sort of rightfully pointing out, it's just early days for us to put these premium additions into our installed base. And as they go through the life cycle of a contract with our company, what we're seeing is more and more of our customers are sort of trading up to these additions as they come up for renewal. So it's not something that will be a big bang immediately overnight is everyone on a premium addition. It is something that as companies have other contracts that they're rolling off from other vendors and now consolidating on the Salesforce. That's the motion that we're kind of running with the premium additions that are there.

And I think it really allows our customers to sort of save money in the process while driving higher performance and higher productivity for every salesperson that they do. By the way, the same is true on the Service Cloud side. You have to have a case management solution, a telephony solution sometimes a chat or channel solution that's out there. And so the same sort of kind of play is what we're engineering on that side, which is more consolidation of extremely with spend into more of a consolidated sales force platform.

So there theoretically will be an upgrade cycle that we could guess at for each of the salesforce clouds over time. That is essentially the strategy you talked about -- the thing that I want to better understand, so this can get pretty confusing pretty quickly because you've got the agent for SKU. You've got what will be the [indiscernible] You've got headless, which is sort of a pick and mix independent of [indiscernible] maybe just help -- you had a beautiful cohort description earlier. -- help us put this into cohort, so we can understand what the different customer...

We're doing a pretty good job of all the names. And so like I said, the naming is something that we're definitely working through as a team, but let me kind of make it really simple. With our agent force addition, you get cloud force, you get agent force, you get data cloud and you get the best of all the Sales Cloud that's there. So we want to have 1 addition that they all come into and [indiscernible] Yes, exactly, a premium tier of all of those offerings.

Now can you buy all of those items discretely? Yes, you can buy all those items discretely but we want to make it an economically viable and attractive offering for companies to get the best of Salesforce in this 1 edition because what we know is that for those organizations that run that agent force 1 addition, their salespeople are more productive. They're using the software more because they have Slack also into that addition. They have our Tableau offering. So they have all the analytics that they need to be productive in their role. And so this is really not just a strategy of sort of loading someone up all the sales force that they could have, it's really we've engineered this from the start to make it the best addition for workers to have the best performance of their lives.

And that is the 60% to 80% or math it back...

That's right. That would be where that number would correlate to.

Okay. Now let's introduce in [indiscernible] qualified into the equation. One of the nuances that we struggle with a little bit is thin in practice sounds a little bit like agent forcing practice but it's not explain the difference?

So I think as sort of the world of agents and agentic platforms are starting to clarify for all of us, it's become very clear that there are many customers that want to have purpose-built specialized agents to perform discrete tests on behalf of their business. Finn is a great example. -- been as a solution that was engineered for the customer service domain, really to help kind of streamline and automate a lot of those routine and interactions that the company has with its customers. And so that is a purposeful agent that is hyper specialized and focused on that area.

Hyper, which is an agent from qualified is actually a hyper focused, purpose-built agent for the marketing domain where Piper's job is to sort of sit as part of your website, enroll someone through your sales and lead qualification process and ultimately try and book meetings with sales professionals. So these are 2 agents in 2 very different tasks on top of kind of commonly what might be a website experience from a company.

Now what agent force does is it sort of orchestrates all of them in the back end. It understands and knows what Fin is doing. It knows and understands what Piper is doing. It knows and understands what agents that are built by developers are doing on this platform so that we know that one agent can start of an experience and then pick it up as a conversation traverses from the marketing world to the selling world, to the servicing world. And so this is where, again, as sort of the world between platforms that orchestrate the journey and agents to sort of power the front end of those experiences, that's how they sort of differentiate from one another.

But it's overly probably reducted to say, is it qualified versus agent force or Piper versus spin. We don't see customers sort of say, I deploy 1 or the other. Oftentimes, what we see our companies deploy all of them. They start with the top of the funnel with the piper, the bottom of the funnel with the fin and agent force sort of orchestrates in between.

And then I also have to budget for anthropic token spend [indiscernible].

Yes. No, not on those offerings. If you're buying kind of cloud as or anthropic as a kind of customer, you do that with them with Anthropic, as part of sort of agent force and part of Piper and as part of FIN, you're not paying sort of for [indiscernible] tokens to make that -- those interactions powered.

Okay. That makes sense. Let me ask the broader question, which is there's an evolution in the software here that goes from being setout based to selling outcomes, selling a unit of work, which we think is quite expansionary for the software TAM where does the budget come from when you're having these conversations?

Yes. Well, I think -- maybe I'll start and Rob, you can kind of add in from your productivity lens. But I think as terms of budget goes, it depends on where the outcomes are. Sometimes outcomes are actually aligned towards savings. So the actual savings become a self-filling proposition for -- for our dollar we save, we can reinvest those dollars in the software that powers those savings. And in other cases, some of the outcomes that we're working on now like in the case of our sales and sales agent domain kind of paying for qualified leads or paying for orders that are actually booked -- and so that's where organizations are happy to invest in those outcome-based agents because the more that those agents perform, the higher revenue that they're actually performing for their organization.

So I would say that the outcome-based agents actually allow us to align value realized with budgetary sort of planning all together. I've not yet met a CFO that says, I'm not willing to give you a dollar if you're actually going to help me make -- so this is an opportunity for us to, like you said, get into more of that expansionary opportunity, especially as our benefits are more jointly aligned with customers on those domains specifically.

I mean I would layer in, we continue to see more and more of the premium for Slack because of the additional work that we're taking on and the economics we're helping our customers with. So we're a little bit different. So when you were talking about, in general, the move to headless and Cloudforce as an example, earlier within Slack, we have something called Slack bot, which is an out-of-the-box agent for knowledge workers that is built on top of Anthropic. And it's a broader sales force level, we think it's important that you'd be able to accomplish the work through whatever your tool of choice might be -- so we want to be able to tie in to any harness that you use, but we are going to have a penned harness out of the box, and that is lockout.

And so when we go to -- when we look at Blackbaud, 1 of the bets that we make is we actually take on the burden of token economics ourselves. And so it's part of the producer per month for our highest-grade plan, our knowledge workers that our customers are able to go to Slack bot and get done much of what they might have done in an open-ended cost tool in another way. And does it mean those other things can go away? No, it doesn't. It means it's an easier way to budget and some of the more specific higher order, like deep knowledge work that needs to happen in those other tools will happen over there. But this is just like a very easy, reliable way for you to give the knowledge work agent to the majority of your employees at a reliable cost and then really target like, yes, they need cloud code or they need cowork or they need an unlimited token spend, et cetera. So but we see often the premium that we're getting coming at the expense of utilization of other tools, frankly. And it's not necessarily elimination of heads.

It's like when you look at Slack, like we'll have a business leader brings Slack into an organization like a sales organization that like, okay, I want to connect Slack to Salesforce, and I have a goal of getting my first sales reps to -- my new sales reps, too productive in 90 days or less. It used to take 6 months, and you connect Salesforce and Slack together, and they can do that, and they're effectively accelerating revenue rather than -- and taking some cost out because they aren't using third parties to come in and train those people. They're actually having the software to do it for them. Or it's like, okay, we're doing a percentage of our internal meetings through huddles now as opposed to these other things that we're doing. So we're taking 30% of the workload off of some of those other tools. Similar things with documents, similar things with agents for knowledge work. So we shave off by being more convenient, economics in other places.

I have 2 big picture software questions. I want to pick [indiscernible]. So the first is, Bill, you've talked in the past about back office versus front office convergence and how agents can perhaps blur the lines. What do you think is the future of what has historically been silos between back and front office.

Yes. I started my career as a production scheduler for Motorola as an intern. And so I was like a back office dude, and then I learned about this magical thing called CRM, which is about like actually working with customers. And so I've always had this like inter rift in my own brain between back office and front office. And what was sort of fascinating is the limits of the technology systems have always been really about the roles that were kind of specialized to perform discrete tests and functions. And when you learn about like a lot of the systems in use for many businesses, those con systems themselves were operationalized or built for the concern of what human labor could do. And so I think it's so fasting right now, and Rob, you talked about and alluded to it with the Slack world, we'll see more agents in our workplace than employees and human employees sometimes in our workplace to perform work in tasks.

And so we have to go back to the ability to reinvent a lot of these fundamental systems and processes and activities that people do because now the technology is going to be always on, always running in our lives really to help us become more productive and to drive higher quality output in the work that we sort of perform. So if anyone here is sort of a back office and front office sort of Maven, I'm sorry to say, I think they'll blend, and I think that they will blend in ways that really break down a lot of categories and a lot of sort of opportunities. But at the end of the day, if we use software to make companies perform better, that's where value systems like Salesforce are really being engineered for from day 1 which has helped make our companies be better companies, help them sort of engage their customers in better ways and help them sort of transform their business, not to recreate yesterday's legacy but really to define what tomorrow sort of looks like.

So I'm very, very bullish about just workplace and workforce transformation in that regard. I never want to have to kind of do MRP scheduling again in my life. And so I'm very, very excited about this ability to fundamentally hand some of that work to agents and to spend energy where I like to spend energy, which is with people and relating in new and exciting ways.

Okay. So the second one is horizontal versus vertical. And you'll have some pretty spiky comments from David Friedberg, gets to the beginning of the earnings call on horizontal leading vertical. You've announced some really interesting life sciences wins. For either of you, what do you think is the future of vertical software?

Maybe I can start Black is inherently more horizontal, I think, than the rest of the Salesforce products. And -- but I think there are opportunities for us to get more vertical. And I think maybe that will be a good segue to Bill. But again, coming back to like the total available market for knowledge workers, like it's in excess of 1 billion people, I think, globally. And so we very much look at the horizontal opportunity for Slack because what's been very interesting is if you look at from the inception of Slack, the challenge that was set forth for the team was, is there a way that we can build software that will make coordination easier to make people -- the hardest part of people's working lives coordination, more simple, pleasant and productive.

And that's a problem space that fortunately has endured pretty tectonic shifts in underlying technologies, whether that be the cloud substrates to now the new models, like coordination is still pain right? I don't think there's a red [indiscernible] article like go read that. And I think there's just -- remains tremendous horizontal opportunity for coordination and orchestration of work for knowledge workers despite everything that's happening with agents that are out there today. I think there are domains that we can verticalize in that we're interested in, you're starting to see us do with Slack code, which is we can -- that span industry vertical. And so I think you'll see us go into stove pipes of LOBs rather than necessarily industry verticals, but I think you expect to see us go long on tech and dev and coding in particular from a verticalization perspective.

Yes. In the spicy commentary, I think, really, the message was about finding out where really your true differentiation in your about to create unique value sort of represents. And if you keep going down that spectrum, there's horizontal and then there's vertical and then there's regional and then there's local and then there's personal sort of opportunities to create value along that sort of spectrum. And I don't think that value stops on any one of those sort of kind of stops along the journey. I do think there's immense opportunity for us to utilize the technology and I think I take a lot of inspiration for what Slack does for me in my life where I get out of my own way, and I work with my team and I work with my group and I work with my organization in ways that I never could do before.

But that just sort of goes on that same spectrum where I'm now working less in a local way but more in a communal way. And I think that as we see this vertical opportunity maybe more clearly, there is great value that we can create making hyper-relevant computing for companies in certain industries that maybe have regulatory requirements and uniqueness that allows us to maybe sort of deliver higher-value services or not. So I'm not sort of a big believer in the versus, I think there's a big spectrum of opportunity there, and we're excited to sort of play in that spectrum broadly speaking.

It's really good stuff. Please join me in thanking Bill and Rob for their time. Thank you [indiscernible].

Good morning, and thank you for joining us for our Q2 FY '27 IR webinar on product adoption and momentum. I am Mark Murphy, EVP of Global Investor Relations. Here with Valmik Desai on the far end, representing the IR team for Salesforce. We're very grateful to be joined by Bill Patterson, President and Chief Commercial Officer, sitting next to Valmik; as well as Connor Marsden, President of Sales and Chief Consumption Officer. These are 2 very dynamic and impactful thought leaders in the industry. Gentlemen, first off, thank you for joining us this morning.

Good morning. Great to be here.

Our goal with this is to address investors' most common questions and topics leading both into and out of our earnings report last week. Those are going to include AI monetization, AI force and other product and partnership announcements, including the one relating to Claude, infusion of AI across the platform. We're going to give you a couple of customer stories to try to demystify what is actually happening on the ground. Before we begin, I want to read you this disclaimer. Some of our comments today may contain forward-looking statements that are subject to risks, uncertainties and assumptions, which could change.

Should any of these risks materialize or should our assumptions prove to be incorrect, actual company results or outcomes could differ materially from these forward-looking statements. A description of these risks, uncertainties and assumptions and other factors that could affect our financial results or outcomes is included in our SEC filings, including our most recent report on Forms 10-K, 10-Q and any other SEC filings. Except as required by law, we do not undertake any responsibility to update these forward-looking statements. So our plan is to start with a brief presentation, and then we're going to jump into your questions. So please, at any time, when anything crosses your mind, feel free to submit your questions in the chat. And with that, I'm going to hand it over to Connor and then Bill.

Fantastic. Thanks, Mark, so much. So I'm so excited to be here today to talk about some of the momentum that we're having in the market today and really helping to turn our customers into Agentic enterprises. And so we've probably shown this slide to you a lot, and I'm going to provide some context on what we see with our customers today. So first, to ground us. As we've been in this Agentic revolution, there's really 4 key aspects to make any agent successful. It's trust, action, agency and interface. And we have rearchitected our entire platform to be able to meet the moment in today's market, starting with trust, trusted context and data that's delivered in a secure way with the right user permissions and of course, with 0 data retention as we start to integrate with the LLMs.

Our action layer, that's our application layer that our customers have used for years and years, service, sales, marketing and analytics. This is where we have our standard workflows, business processes or custom workflows and business processes to meet the unique needs of our customers. For example, Tableau today has 33 million semantic layers built in that we can now extend into our layer of agency. And you're probably seeing some new names on the list, Casey and Hunter and Marshall.

As we continue to extend our Agentic footprint, we now have out-of-the-box use cases that our customers are leveraging to drive our Agentic footprint. So for example, Casey. Casey takes advantage of the best of our Service Cloud application and the best of our data footprint underneath to provide great resolution for our customers with no human interaction. But when you do need to pass it over to a human, it's passes the context of that conversation so that human can pick up with their own agent to help solve that customer problem. And all of this is extended into our headless interfaces.

Now whether that be our traditional Lightning interface or Coworker, which is our fastest-growing AI product that we've had to date or Slack or Teams or this great announcement that we made with Anthropic with Claudeforce, being able to take the best of Claude to create custom interfaces that combine the best of Salesforce and the best of other applications into dashboards, analytics and purpose-built views into your data to help people get work done faster.

A great example of this is SharkNinja. SharkNinja is a great commerce customer, service customer, Data Cloud customer today on the platform. They launched a service agent together with our shopper agent and saw a 6% increase in conversion. But then they wanted to build their own custom agent, an unboxing agent. And this unboxing agent when you opened up one of their espresso machines would help walk a customer through how to set up the machine, how to configure it, how to make your first espresso, how to order any consumables for that item. And it required each layer. It required the agency layer. It required the service layer and it required the trust layer with Data Cloud, all uniquely positioned.

And they saw a 93% resolution rate for the customers who use that agent and only 7% had to be promoted to a human. So that represented a better customer experience and significant cost savings. And what's unique about our platform today is that when we look at the data lakes and the hyperscalers, they may have a data layer and they're building out an agency layer, but they don't have an application layer today. Or if you look at the frontier models, they certainly have an agency layer that they're building, but they don't have an application or a data layer and really understanding the metadata and the context and the semantics of the data in which they're working on.

And the traditional application providers, they may be building an agency layer on top, but they don't have the robust data layer that we have underneath. We are uniquely positioned in that we have all 4. And all the providers wish that they had Slack today to drive AI together with your employees and the marketplace today. And through this investment, we're seeing an absolute -- our Agentic footprint rapidly expand. And so when I started as the Chief Consumption Officer in February to make sure our customers are getting value off of their deployments, it was really about Slackbot, our , which is our internal agents and Agentforce, which was the foundation for our external agents that we are providing into the marketplace.

In H1, through both organic and inorganic innovation, we brought Piper and Marshall, our Momentum acquisition for conversational AI for our salespeople. And then Coworker, which we launched at the tail end of H1. And that's going to rapidly expand into H2 with our pending Fin and Contentful acquisitions, our Marshall, Albert, which is going to be taking the best of SlackBot, but extending it to non-Slack environments. But then Hunter, Casey, Page, Carter, just to name a few, and then all powered with our headless and Claude applications.

Now what's really interesting is we have 10,000-plus customers today that are uniquely using one of our AI products. And now we're seeing significant momentum as we're having more success of people adding a second, third and fourth AI solution from Salesforce. So we can continue to sell into our base and driving new customer acquisition, which roughly doubled since the beginning of February, the number of customers who had our AI products in production. But now we can start to cross-sell into that base of customers to show more value from an Agentic standpoint from Salesforce. And what this really has helped us lead to is fantastic financial results.

Our Agentforce and data ARR grew by 200-plus percent to $3.9 billion. Our Agentforce ARR is up 200% to $1.5 billion. And most impressively, and this is a measure of value and the work being done on the platform, we saw a 97% increase of our AWUs to 7 billion. Now we get asked the question all the time, like who's using our platform? And is it old industry, new industries? 9 out of the top 10 AI-native companies have chosen Salesforce to be their central nervous system in their brain for running their operations today. Now what's another thing that's enabled some of the success? So one of the things we talk a lot about here at Salesforce is speed to value with our customers. And we've made a lot of really smart investments to really unlock the value for our customers.

So starting on the right-hand side of the screen are builders. Builders are our version of FTEs. We have roughly 600 today in the market that are all embedded with inside of our sales team, working hands on board with our keyboard, and we're going to be more than doubling this investment towards the end of the year. We really saw the value in out-of-the-box agents. That's the Casey, Hunter, Piper because when you have an out-of-the-box agent together with our apps and our data layer, we're really able to drive fast time to value and get our agents up and running in 30 to 45 days. And then we wanted to unlock the entire developer community. And so we launched Agent Script, which allows us to plug in Agentforce into Claude or Codex so that they can use the development environment of their choice and drive deployment with inside of Agentforce. And that's why we're seeing the time to deploy significantly improve.

And then just one other aspect I want to talk about. All these since we've identified every layer of our platform for our traditional applications, sales and service, we're able to use these coding tools, Claude and Codex to actually configure our Salesforce environment. And we're seeing a 40% improvement on how quickly we can get our customers onto our platform. One large major global retailer that we are working with needed to do a contact deployment with agents on the front end. Their internal engineering team gave them a quote that it was going to be 35 weeks to deploy a custom build of that solution.

They chose Salesforce, we were able to get it live in 6 weeks. It's phenomenal about how we can use the tools to deploy our platform to bring agents along for the ride and really turn them into an Agentic enterprise. And this is really reflected in the new customer acquisition that we've been able to drive to that. I briefly touched on SharkNinja. We have Dell that's using us for supply chain today, 19,000 individuals at Dell uses to route complex supply chain requests, automate the processes, and it's saving roughly 30 hours per week per team. We talk about fast time to value, Live Nation. They just had a great festival, BottleRock up in Napa Valley. And we were able to get an agent up and live in 30 days where they had 37,000 guest interactions, and they're going to be expanding this to all their festivals across the world.

And then Wyndham. Wyndham deployed an agent in their contact center that saw a 25% increase in the average -- or decrease in the average handle time that they were able to deploy. And that required their application layer with agents and then on the foundation of data that we're able to bring to bear. So we're seeing fantastic momentum today with our platform, with our customers. We're seeing expanded usage of our apps. And now we're starting to see customers go back to the well again and add the second, third and fourth Agentic solution from Salesforce. And this is going to drive monetization at every layer of our platform. And so now I'm going to hand it over to Bill to maybe talk about how we're trying to innovate from a pricing perspective.

Yes. Thank you, Connor. And look, I think one of the reasons that I love working at Salesforce, it's the slide that you see here. It's all the customers that we're bringing really into this Agentic era of truly transforming their business. But I can guarantee you if you have the wrong pricing model and the wrong monetization model, they're not going to transform. In fact, they're going to be really confused. And so one of the things my team has been deeply focused on is really simplifying and streamlining a lot of our pricing and packaging opportunities really to give customers benefit at this time. Enterprise software has always been complicated. It's always been challenging to sort of price by feature or priced by size of an offering or how much scale do you want to put into your solutions.

And so when we had this moment really about helping our customers, we went back to sort of the foundations of Salesforce and said, what we really need to do at this time is invent a new pricing structure that really aligns to the benefits that customers realize from this new technology that you just went through. And that's where, number one, the first area of focus for my organization is ultimately helping with better predictability, but really better value realization for customers. And so we've invented a whole new model called outcome-based pricing.

Our help agents like Casey, like you talked about, are priced by the resolutions that they drive for our customers. And if they don't resolve the issue, you don't pay for the offering. And so this is truly putting our money where our mouth is around driving better value realization for customers who utilize our technologies. In addition, it's not always the case where you can have price agents on an outcome-based way because sometimes agents work across multiple disciplines, multiple domains, multiple offerings. And so what we've also done is really gone back to the drawing board around creating better value-based bundles and flex credits to flex and scale and stretch the platform as needed to really utilize and deploy the capabilities you need to make your business successful.

As Connor started, the architecture is so pervasive now and it's so impressive and it's accelerating every day. But really the flex credit manifestation lets you use all of those capabilities to use what's right for your business and just give you kind of more value at scale. And what we learned at this moment in time is customers are on different sort of modes and pacing of this journey. And so we need to ultimately be more flexible. We need to kind of offer new ways to sort of activate your business on Salesforce. And so this is sometimes where you utilize things like pay as you go. So I just need to pay for enough if I'm consuming from the platform and put it into action or thank you. I use kind of deeper contract structures like our Agentic Enterprise Licensing Agreement, which gives you unlimited access to all the technology, utilize what you need, put it into kind of practice and then scale that across your business like never before.

So these new structures, more predictability, more customer value, more flexibility are fundamentally transforming how we serve our customers in this moment because at the end of the day, if it's confusing and if we're not serving our customers, it really doesn't matter. Let me build a little bit on behind the scenes about how we're kind of using this new Headless 360 monetization structure. We got a lot of questions about this from developers and a lot of our partners. But essentially, as I mentioned, everything is a compounding model here at Salesforce. And so Flex credits serve as the foundation of every sort of consumption utility that we have in our offering. And so for -- if you just want to sort of get started, you use what we call our Salesforce Foundations platform and Flex credits, and that just sort of runs the meter for what you kind of access on the platform itself.

One of the things we're really excited to do, though, is to offer kind of more of that predictability, more of that value for every user. And so we've created what we call this Headless Add-on, which packages a lot of work capacity for every user, make it easy. So maybe you and I use the software a little bit differently. You're probably a lot more intense on the software with how much kind of consumption you're driving. But essentially, that headless add-on monetizes just in a simple way for every user to access the platform, and it gives you enough capacity that you need to get your job done.

So we take this Headless Add-on, put a lot of more capacity. This is all the AWUs you need to be productive. And now that becomes a simple way for every organization to start accessing new technologies like Claudeforce, the partnership we just announced with Anthropic. Now we're also going to take that Headless Add-on and package it into our premium edition. So as you move up the value stack for Agentforce for sales or Agentforce for service or maybe one of our Agentforce for industry offerings, this capability kind of nests inside of that sort of productivity suite. And so this is where we start to see the magic start to occur. All of that AI centralized and organized by role, optimized for every user makes our users very productive with the amount of capacity they need and it's really that focused productivity they require to get their job done. And so that Agentforce service or sales add-on also next in our premium edition.

So this is where this compounding monetization simple model, it just embeds into every prior edition embeds into the top tier. This makes it really simple for our customers to plan, predict and scale their use of Salesforce in this moment in time. And ultimately, the Agentforce 1 Edition is the best of Salesforce for all roles. It's what includes our traditional applications like sales and service. It packages Slack, it packages Tableau. And so ultimately, this is where the best of all of our company's offerings come together in that Agentforce 1 Edition. And this is really exciting to see the growth of this offering because it means more customers are finding more value with the platform in this moment in time. And ultimately, at the end of the day, it's about transforming their businesses with the technology we serve them with.

I know, Bill, when I go and talk with our customers today, and there's a lot of concern over the cost of AI. We've heard some of the stories on Token Max and things of that nature. We have the flexibility to provide a limited agreement so we can take that -- we can provide a predictable model on what the cost can be. If customers are using multiple solutions from Salesforce, you don't need to precisely predict what the consumption or anything is because you have flex credits that are fungible that can be plugged into your data solution if you end up using more of our data products or across your various agents. And so we can kind of meet the moment for wherever our customer is on the journey to provide them the right contracting and pricing to meet their requirements. That's where our breadth and scale really comes to provide comfort for our customers that they know they're going to only pay for what they're going to be using and whatever iteration that we decide to package it out there.

I was with a customer just recently, and they said their spending is up 30%. And when they actually did the math, their productivity was up by about 3%. And so there's just a fundamental mismatch today between what is going on in the spending environment and then the value realization environment. Our mission is to change that. It's to be easier, to be more simple, to be more valuable for every customer. So you're not counting how many tokens, you're counting how many leads that you process or how many new orders that you sort of manage or how many cases you resolved on our platform.

And then just the last thing that has been eye-opening for me personally is users. What this model is providing, especially with Headless is it's actually giving us access to more users internally, people who want to take advantage and consume the application. And it may not be the full Salesforce application. It may be a headless application that they're engaging with and consuming. And so it's actually expanding our TAM across the addressable user base inside of our customers.

Yes. I think serving customers is not limited to one department. And so the opportunity to really use this new platform, this new capability to reach into -- across the organization so that we all can work in service of our business and all work in service of our customers, it's ultimately where we're seeing a lot more value be created and a lot more sort of growth opportunity for our customers to realize on the basis of what this platform can do for them.

Yes, it really unlocks more -- it just really unlocks the system. So if someone in finance needs to address a service or a pricing issue, they can get access to the right screen to get their work done efficiently with AI guiding their journey.

Absolutely.

Thank you so much, Connor and Bill, for all those comments. We'd now like to go ahead and get into the Q&A session, and we've left plenty of time for that. [Operator Instructions] While we queue up those questions, I want to begin with the first question we see, which says on Agentforce and Data 360 pricing, you've iterated through several models, is the pricing approach getting clearer for customers? And are you seeing their preferred way to consume it become clearer too?

Yes. First off, thank you for the question for those that submitted online. First off, is the pricing getting more clear? God, I hope so. We have actually kind of been iterating along the way around different moments of what the AI can do. And in some cases, like you've seen, where Agentforce started as a platform where we're building a lot of capabilities and it was about building agents, we have to price in a platform kind of way. And so that really made it so you were probably maybe too atomic for what you were monetizing in the offering.

As we have these new worlds of out-of-box agents or complete agents, finished agents to serve a domain that can get to an outcome, it's a lot more easy to predict and scale those kind of price and monetization moments because the value exchange is clear. And so as the feedback continues to come in for our ecosystem, it is getting more clear. It is getting easier to deploy. It is getting easier for customers to sort of realize value. And that value exchange is ultimately kind of where, again, as Connor and I were talking, very different from sort of how many tokens do you need to be productive. So I definitely think there's a much more clear approach. It's not limited Agentforce. We've also been sort of simplifying our approach with Data Cloud and sort of some of those underlying technologies to serve the agents with. And so as I mentioned, we sure hope it's getting clear. Our customer feedback has been quite strong. But again, how that measures and manifests is the more that customers adopt, that's where we see success.

It's natural when you have a kind of Cambrian explosion in all of these new technologies, the new surfaces, the new form factors there has to be experimentation in the early innings, right?

Well, we always say the inventor of the ship also invented the shipwreck. And so we really needed to make sure that we went through these iterations and tested what works, what scales. But ultimately, at the end of the day, our mission isn't just sort of put technology in the world and hope you get value from it. No, this is where we go back to our roots in Salesforce and say, you get value from the offering and get value from the agents and the value exchange between our customers and our company is clear. And so yes, I definitely appreciate the feedback. It's been a lot of iteration, but I think the signs for customer benefit and growth of usage, as you've seen, Connor, are starting to show the signs of a much more clear approach here.

Wonderful. Thank you. Our second question, I should read the names as they're coming in, is coming from Luke Heinen of Madison Investments. Thank you, Luke. And this says, how are you or will you monetize Slackbot if customers prefer to use Claude within Slack, is there a way to monetize that? Or does that limit the Slack revenue opportunity?

Yes. Great question. Thank you, Luke, for the question. We are planning to monetize Slackbot as part of your Slack subscription. First off, Slack is a per seat or per user subscription access fee. So everyone who sort of wants to access Slack pays a per user per month offering there. Inside of that per user per month offering, we give you a lot of capacity to use Slackbot to be productive. So if you're asking questions about your customers, you're asking questions about sort of your suppliers, you can use Slackbot for all kinds of sort of productivity experiences. And every user gets a lot of questions that they can ask.

Now if you have power users and you go beyond that sort of inherent supply, you can always add more capacity for Slackbot to do more work for you. So again, this is where we load you with enough work to get started. You utilize what we call our Flex credits to expand that usage capacity as you sort of need -- usage grows over time. And it really allows the organizations to sort of kind of crawl and walk them truly run with AI into their offering.

The second part of your question about embedding Anthropic inside of Slack, does that limit Slack's revenue appeal? Absolutely not. Anthropic is good at some things. Slackbot is good at some things. Other partner offerings and other AI offerings are good at some things. And so what you really should expect from Salesforce and ultimately, with Slack as a surface in the canvas for us is expanding the number of opportunities that other productivity systems can integrate because what we know is integrating in the tools of choice of the customers that we serve is going to ultimately make them more productive on the platform. So we don't see that as a limiter. We actually see it as an amplifier. And we do think that it also drives happier users. By the way, happier users drive more usage, more usage drives more retention. So this flywheel of opportunity and growth truly starts to get groundswell and steam.

So I actually had this question happen with the customer where they were using Claude and they're evaluating Slackbot. And fundamentally, Slackbot uses Claude as the underpinning of the LLM. But we provide practically unlimited capability, unlimited use of Claude with inside the Slackbot licensing model today. And so what I told the customer is I said, "Hey, you can use Claude and pay per token or you can have an unlimited use of the application inside of Slackbot and not worry about your token expense, have the same capabilities, have the same ability to share skills to be productive and drive it. And they found that to be a much more appealing model than using Claude from a token perspective today, and they opted for SlackBot.

And so this is, once again, from a pricing packaging standpoint, what are the scenarios where you're going to uncap the amount of use of the application spread it across a lot of users, we're going to have power users and light users and have the right model versus a scenario where you're going to be paying per token that potentially can expand where you're not getting the productivity and the customer opted for Slackbot. And that's a pattern that I see playing out over and over again.

Absolutely. If I could add 2 more points on the Slack piece because it's so important to the core strategy. The Slack interface is expanding as we speak. The innovation that the team is driving there. We've talked about Slack Code briefly at earnings. When you come to Dreamforce, you're going to see a full version of what Slack Code looks like for a developer. That interface, that UI, where you're able to get your work done where previously you might have been switching between 4 or 5 different screens just to get that workflow complete, Slack is now becoming that hub for a lot of that work.

Two other stats that I think were really impressive from the quarter. First, Slack had its strongest net new AOV performance since we closed the acquisition in Q2. And second, since we launched Slackbot, just in the period from GA to today, we've seen a tripling in upgrades. So when we think about the value of Slack and how we're thinking about this broader interface layer, the more we integrate with more partners, bringing those amazing multiplayer agent experiences into that interface, the more we drive more Slackbot usage consumption adoption, the better for Slack's overall business performance and frankly, the better for how we're actually interacting with Salesforce, right?

For me, I do a lot of customer meetings when I get asked to, I love to do it. Previously, someone from Connor's team would come over and be like, okay, let me give you the prep. Let me pull stuff from our Salesforce org. Let me give you the full view. I don't have to do that anymore. I just ask Slackbot, what should I know about X customer? What's their recent products? Are they having any customer success issues? What's the team trying to sell? And it gives me a full prep doc directly within my Slack channel. And within that prep doc, I have Connor's team coming in and telling me, "Hey, here's where we really want you to focus." So this flow of the interface being everything that you would have normally done across multiple different layers showing up in Slack, we think is super, super exciting.

It's a game changer. It's addictive. I can tell you that as a new joiner that never had access to Slack and had never seen it. I said recently, if you take it away from me, you're going to have to try it out of my cold dead hands. That is how addictive that product is. Thank you for that. Next question is going to be from Alex Zukin of Wolfe. How should we think about how much certain Agentic workflows cost? Is there a dollar value per lead? Or is there a CPQ agent or a marketing agent? How many AWUs will that burn? And how should customers think about fitting this into budgets, particularly with something like Slack and code channels, et cetera? Thank you for the 3-part question there, Alex.

Alex, thank you for the thoroughness of the question. Look, let me tell you how our team thinks of setting pricing in this moment because as you've seen and heard earlier, we went through a lot of iterations on this topic. How we think about how much certain Agentic workflow costs is really about setting a value and a value for how much -- how valuable is the offering that we're sort of putting into the hands of our customers. For example, we know there is a dollar per lead because that's ultimately how organizations size and scale their own performance and their own planning and budgeting activities for how many marketing qualified leads they need to really sort of enlist and fuel their sales funnels, if you will.

So we do look at this sort of in terms of what organizations today pay, not just in software, but total cost, labor cost, workflow costs, software costs, the integration costs. All of that sort of equates to some degree of utilitary budget that an organization would sort of plan for. What we're doing is we're competing on the basis that we think we can make that easier for organizations to do based when you use our technologies. And so I don't actually want you to be think -- the second part of your question about AWUs that burn, I actually don't want you to think about that at all.

I think that's an internal measure of utilization for how we are sort of seeing how much performance happens out of our platform called the Agentic Work Unit. But what I want you to be thinking about and what I want your customers and the industry to be thinking about is really kind of the labor and the economic offset of it used to cost me maybe $5 to do a phone call. Well, now I can actually do it for $1, $2 per resolution here on the Salesforce platform. So this is where we want to just think about kind of offsetting kind of expenses for our customers and making it much more economical, much more easy to scale using the benefit of our platform that can now serve that unit of work.

Next question is from Kirk Materne of Evercore. Can you talk about how the horizontal pricing strategy for AI complements your industry cloud pricing in any industries where these new AI pricing configurations are landing best?

Yes. This is fully a pricing webinar today, I think. The horizontal pricing strategy and the industry pricing strategy sort of follows that same structure of compounding value like we talked about. where the horizontal platforms are priced sort of as a baseline and the industries and the vertical offerings that we have where they have much more acute value and much more precise value, therefore, higher value, if you will, are priced at a premium offering for our customers. And so in terms of sort of that compounding value in that stair-stepping model, our Financial Services Cloud is priced as an uplift on top of our core Sales Cloud, if you will.

And so this is where you see sort of that value compounding sort of structure really come to fruition. Are there industries where AI pricing configurations are lending the best? I can give you 2, and then maybe, Connor, you can add in. I think financial services is an area specifically where the data sets the regulated sort of environments, the sort of out-of-box functionality we built for kind of KYC and household management, these are specific areas where we have price premium capabilities for sort of the financial services industry that we price on top of our core Sales Cloud, but which maybe doesn't need those areas in the horizontal selling kind of way.

Same thing on health. And again, where you see the sort of regulated environment where you have PII information, you have information around HIPAA compliance. Our offerings that have these sort of compliant boundaries have all of this nested value on top of the horizontal offerings. And so this is where you see sort of AI for industries really command that premium price point, if you will. Maybe you can add a few from your standpoint.

I mean we're seeing a lot of momentum in our consumer goods market right now with our CG cloud. And we're seeing all the leading CG players migrate over to our platform. And there's really 2 scenarios. There's the customers who already had CG Cloud today and now they're identifying their processes with our out-of-the-box agents for CG Cloud, and we're getting an uplift as they're starting to deploy those agents. And we're seeing a lot of net new customers that are buying, deploying and creating global models for CG Cloud, but they're starting -- in the starting position, they're identifying their work processes.

And with CG Cloud, the out-of-the-box use case, they're able to deploy it in 30, 45 days on top of their existing platform. And so we're seeing tremendous momentum for that right now because the AI is purpose-built for trade promotion management, for retail execution. It's an agent that sits alongside the salesperson to understand what's the package that I sold to the customer? Have we sent the rebate check back to them for their -- for the display that they put inside of their store. All the approvals and workflows are flow through the back end of the system that the agent drives it. And it brings in the best of both worlds, the determinism and the probabilism on top of the industry solution. So we are seeing these industry models and these industry purpose-built agents really driving value and speed for our customers, which is so critical as things get more competitive in the marketplace.

Wonderful. Our next question is going to be from Samik Chatterjee of JPMorgan. What are you seeing in the early days for adoption of Agentforce 1 Edition? And do you envision the -- how do you envision the customer sales motion to upsell working?

We are seeing an explosion of upgrades to the Agentforce 1 Edition. It's been absolutely a phenomenal motion for us. And we think it's going to accelerate because we're adding Headless into the model as well. And so before, it was about unlimited Agentforce for the internal users. So an agent sitting alongside a contact center worker to help walk through a customer issue and to guide them through resolution, leveraging the full knowledge of the organization. And now it's going to be not only that agent sitting on top of, but then also what are the headless interfaces that, that employee is going to potentially want to leverage to drive. And so Agentforce 1 Edition has been a big winner for us, and we see it accelerating as we move through the foreseeable future.

Yes. I'm so excited about the Agentforce 1 Edition. When we sort of invented that offering for our customers, we were listening deeply about their needs were. And what they all told us is, number one, they wanted to be multi-cloud customers. And so what Agentforce 1 Edition has is that it not only has -- let's take in the case of Sales Cloud, has all the best offerings for what Sales Cloud can do for a sales organization, but it also has Slack. It also has Tableau. It has Agentforce. It has Data Cloud and it has Flex Credit sort of in its core. And so what this means is, for the first time, Sales Cloud customers are experiencing the benefit of Slack or the benefit of Tableau or the benefit of Data Cloud or Agentforce, et cetera. And so what we're really so excited about around sort of the sales motion side of this, a lot of it is user-driven. A lot of it is sort of application-driven.

It's almost like product-led in that more usage, more scenarios of expansion now start to occur because all of those offerings are natively there and available for the user to start discovering. And I think your point is dead on, Connor, around Headless. Headless really allows usage walls to break down between applications. It's just about getting your job done, and we're serving that in new and exciting ways with the platform itself. So I think the sales motion side, the adoption side is clear, but the selling motion is a lot of user-led expansion and user-led sort of iterations on what they do with the software every day.

That's right.

Our next question is coming from Paul Oppenheim of Ardsley Partners. Slack as the front end for usage is a really powerful position. Given that position, can you monetize carrying various frontier models similar to an Amazon Bedrock. So who wants to take that, I guess, the notion of the model router or pick up?

Well, maybe I'll start from the monetization lens and maybe, Connor, you can kind of expand on that from the customer lens. But like, look, I think this is a moment where customers want choice to sort of complete their work. And I think this is also a moment where some models are very expensive to complete low-value work. Some models are very cheap, but maybe they're not as efficient for all the kind of work exercises that are required. And so I think our position continues to be that with Slack, and we agree with your sentiment, by the way, that Slack is probably the world's best canvas to consume models with.

We think it breaks down the sort of the technical digerati of making it hard for the average user to sort of find AI productivity in their everyday work and workflow. And so we're really excited about Slack at that front-end kind of experience. On what kind of work can you get done? Yes, we think customers have the right to choose. We think customers have a right to choose various frontier models to drive various tasks. We're very, very excited about the partnership we just announced with Anthropic. We're going to expand out those partnerships in new and exciting ways to really be focused on the right tools for the job. It's not a really healthy strategy for you to sort of cut butter in the enterprise with the chainsaw. And so what we really want to do is make it easy for users to get productive and use the right model to serve the right task into the enterprise. You can probably expand that from what customers are telling you.

Well, I mean, customers are really excited about Slack because there are so many unstructured conversations that will happen about a customer. And all the frontier models want access to that data because they all see the value that taking their frontier models, applying into that data, the unique scenarios that we can start to create for customers, together with the Slack data, with the Salesforce data, with the third-party data that we're ingesting into Data 360 to get the complete picture of the customer. And so it really is the ultimate layer of bringing all the pieces together.

And clearly, like the more models we can put forth, the more choice we can provide, cost options and whatnot, what's the right model for the right job, that's definitely an area that I think we're going to continue to drive and explore because ultimately, it's going to drive more value for our customers. And the Frontier models want to get into it because it demonstrates -- it's the best demonstration of the value that their models can bring to an organization to really pull that full knowledge together and solve that problem on behalf of an employee or customer.

Everyone is aligned to Slack as that key surface. Next question we'll take is from Ted Wang, ExodusPoint Capital. Can you talk more about the industry cloud opportunities in particular, Life Sciences Cloud, wondering what the growth opportunity is there and for other industry clouds?

Yes. I think every customer we serve is in an industry. And the good news is we have a lot of industry sort of prebuilt capabilities to transform and help those organizations adopt sort of AI like never before. We have 13 different industry solutions that list continues to expand, of which Life Sciences Cloud is one of our newest ones. We are so excited about this opportunity because every organization is reinventing itself, not just on the basis of what human capital and human labor can do, but now this expanding opportunity around agentic labor and agentic sort of kind of process enablement gives us an amazing sort of runway to keep helping every one of these customers transform.

The Life Sciences Cloud opportunity, and we've seen this really accelerate since putting it into the market roughly, I think, it was less than a year has been one of our fastest-growing industry clouds. And I think because of this opportunity to safely kind of transform with the Salesforce platform, a trusted platform like you showed in the architecture today, that is what gives a lot of companies sort of the excitement and trust that they should transform their industry with Salesforce. But also these prebuilt workflows, these prebuilt capabilities, these Agentic sort of front ends give us new ways to sort of help these customers sort of find new value creation opportunities for themselves.

So this, I think, cloud continues to have huge runway ahead of it. I'm very, very excited about this customer sort of transforming. I've been with a lot of global companies, especially in France that have really taken this to cloud and embraced it to accelerate. I think this is going to be one of the growth drivers for our future on the industry platform at large.

So similar to CG Cloud, Life Sciences is going through a big upgrade cycle right now. And for the most part, we weren't in the Life Sciences business. So this is all greenfield for us. And so now a customer is trying to sit down, and I made this in my opening comments, the application companies today, they have applications, they may be trying to build some agents. They don't have the data layer, they don't have the interface layer. And when you think about life sciences cloud to deploy it to satisfy the needs of the end sales users, you want to have the right interface to meet their -- you want to have the right interface to meet those user requirements. You want to have agents built in.

You want to be able to leverage the full totality of all your knowledge internally. And so to some degree, we're bringing in a full platform that's fully identified with data, agency workflow processes and interfaces that we can in a greenfield space where we didn't have a solution, and now we're selling the full picture. We're not just selling an application from that perspective. So to some degree, it's an unfair advantage that we have right now to capture share inside of life sciences and these other industry clouds built on the standard industry framework that we're building for these industries that gives us a real advantage, and that's why we're seeing such aggressive revenue growth in life sciences and CG and the other clouds that we brought to market today.

We need an agent that can manage microphone failures. That's one thing that would help us. All right. Thank you, Ted. The next question is from Jackson Ader of KeyBanc. Are there sales enablement changes or product training initiatives that need to be undertaken for the new AI motion? Or is it pretty seamless for the existing AEs?

Yes. I'll take this one since I support a large part of our sales organization. Yes, there's a change. And it's a change in conversation. It's a change in where value is being derived with inside of our customers. And we have a very robust enablement motion that Miguel Milano has been driving of training and driving. But the best training happens on the job. So we're not just talking about AI solutions. We have them deployed internally. So our employees are using Slackbot on a daily basis. They're using our Hunter agent and consuming leads from our Hunter agent inside. If you go to salesforce.com, you'll see Piper, our Piper agent that helps generate opportunities that helps create custom presentations to customers as they're looking for product and then booking customer appointments.

So our AEs are living the experience at the same time as they're selling the experience, and that makes it so much easier for them to translate. When I go in front of a customer today, I used to spend -- my teams would spend 1 to 2 weeks putting together a background and a brief on what conversation to have. Today, I just type into Slackbot, Hey, I'm meeting with the CEO of XYZ company, what should I talk about? And it goes into Slack and it sees all the unstructured conversations. It goes into Salesforce and it says, "Hey, here's what they bought." It goes into Service Cloud and says, here's the issues and the cases that they have. And so I get the complete picture of the customer, something that literally has saved weeks of time for our customers today, and our AEs are using those tools on a daily basis.

So yes, there's training. There's new business value models that we're creating that we have to train our end users up. But the most important fundamental fact is they're living the experience on a daily basis, and they're able to bring that experience to our customers. And when our customers see how we're using AI today, it's very eye-opening, and it drives a lot of momentum and accelerated sales cycles as a result.

Yes, Connor, you and I have worked together for a long time, even time before our time at Salesforce, we're getting a little bit old now. But the transformation that's going on in the selling environment, I think you personally need to take a bow on because you are leading our company by example around what does it mean to really drive utilization and activation of everything that has been sold so that you can ultimately drive value. And this is not just me your friend saying it, I think you actually are leading the industry in this way. The element of transformation that's really exciting is when we grew up in our careers, you used to sell first and then sort of deploy.

Now what's happening is organizations are deploying, seeing that value and then selling is almost happening on the other end of that, right? So it's almost a complete inversion for how we grew up in our careers together for what selling kind of means. And so back to your point, it is not exactly the easiest transformation. We have to almost unlearn what we've learned for 25 years around sell and then deploy. Well, now it's about a deploy and sell and realize value kind of world. And I think that ultimately, this means that we will add more value for the customers that we jointly find a lot of passion serving because customers are actually getting benefit from the software, not just getting more software.

Yes. No. I mean -- and that's why we've been really aggressive of putting consumption into our frontline sellers' commissions. So if they sell something, they have to deploy it. And that's going to be tied to part of their compensation. We're going to lean more fully into that because we all know that if customers are getting value, they're going to be more asked to buy that second and third application. Going back to my original comments, we have 10,000-plus customers using our AI solutions, and now we're seeing the second, third and fourth because the customers, when they see value and they see value being created, they want more of that because that's going to translate to them supporting their customers in a deeper way.

Yes. And I think that point is actually being crystallized in one of the metrics we share with you regularly that 50% of the Agentforce, the agent-specific use case, 50% of the bookings there is actually coming from customers refilling in that exact motion, right? So I think when you get the salespeople aligned, the customers aligned, you deploy together, you get value out of it, they're willing to come back and buy a lot more.

Yes, well said. Wonderful. Our next question. This appears to be coming from a customer. Andrew Russo of BACA Systems. Thank you, Andrew. We have Agentforce 1 Edition. We upgraded to it to be able to drive business value for our team. Users are loving it. We are using Salesforce across our entire organization from sales to the manufacturing floor and finance. That being said, will Salesforce be evaluating moving sales and other automated things like sales autonomous sales agents to be outcome-based. It is much more palatable for our CFO to look at trading payroll for outcomes versus payroll for usage.

First off, we always listen to our customers. Andrew is someone from our customer community we deeply listen to because he deeply cares and is routinely at all of our events and gives this feedback to us live. Andrew, first off, I'm happy to tell you that, yes, we will be driving more outcome-based sort of offerings in sales. We'll be driving more outcome-based offerings in service. We'll be doing more outcome-based offerings across our platform because we 100% agree with you, my friend, that the more that we can drive value exchange between us and the customers that we serve and making it easy for you to talk to your CFO, we want to make that easy for you because, again, tracking tokens and token spend doesn't always equate to business value exchange.

Outcomes really allow you to sort of start to scale and predict the value, not just sort of consumed, but the value realized from our platform. And so this is something that we're really excited to talk about at Dreamforce in just a few weeks' time. You should expect to hear more about as more of our agents move into the outcome world, how they will find mutual value between us and your organization. But again, thank you for everything you do, Andrew, for our community, and thanks for coming on to the webinar here this morning.

Yes. And let me just add that this is where our scale really comes to advantage because we can have various pricing models for our customers based on the individual use case versus a collection of use cases, whether it be agents or data or Slack, from that standpoint. And so charging on a per lead that gets qualified through Hunter, that's a direction that we're definitely going. Now clearly, from an AI standpoint, when I see sales -- autonomous sales agent, like how far we can push sales into agents to take that off of the sales team. So smaller transactions just go through an autonomous agent, we're going to be pushing that as hard as possible. And in fact, internally, we're actually seeing our agents drive revenue in smaller transactions with no human touch.

So as the models become more sophisticated, as our determinism becomes more effective, we're going to push down into that. And I think that is going to create new monetization models for Salesforce and for our customers, so they're only paying for the value that they are going to drive.

Thank you, Andrew, for those thoughts. Our next question is from Tyler Radke of Citi. You mentioned last quarter that a large portion of Agentforce net new AOV was driven by top-up credits. What are the high-value use cases that are driving top-up actions? Does that change with the Claudeforce announcement? And maybe one of -- maybe for the benefit of anyone who's not familiar, just explain what a top-up action is.

Sure. Essentially, when we have a sale of credits to a customer, it's -- you're buying capacity for an Agentforce or a Data Cloud to be used for kind of usage that might be a service use case or a sales use case or a data kind of processing use case, et cetera. And so an initial sale might supply you with a certain amount of capacity. When we talk about a top-up credit, that means a customer has consumed all of that capacity and needs to refill the tank and put more in the tank because their usage is sort of kind of going beyond maybe their expectations. It is a great sort of bellwether of success that customers are actually driving sort of usage and retiring the credit that they sort of bought against one of the usage scenarios because ultimately, as we've sort of mentioned earlier on outcome-based pricing, that means they're getting benefit from that usage.

So of the high-value use case that we see for driving top-up actions, one of the clearest signs of value between us and our customers for Agentforce specifically has been in the customer service domain. Today, we use Agent force at Salesforce. You can go to help.salesforce.com, see the solution live and running on our own website. You can see how many cases we're sort of driving, how many sort of interactions that we're resolving all autonomously with the power of that platform. That is the kind of clear use cases that are truly starting to emerge is customer service and customer service resolutions are probably the highest value use case that's there.

There are others. We also have deployed Piper, our kind of digital SDR agent on salesforce.com as well. You can have a conversation with Piper. Piper actually is an agent that is working to qualify you, nurture you, put you into a selling funnel. That's the kind of usage that again, you buy a bunch of credits. Once you kind of get more usage, you top up those credits because you're obviously qualifying or nurturing more leads using the power of digital agents sort of to drive that activity. What are some of the others that your customers are seeing?

Yes. So it's a great question. So from a consumption standpoint, the service use case drives roughly 5x the amount of AWUs as the other use cases just because of the intensive workflows and the sophistication of the determinism behind the agent that's required to satisfy. That's why we're so excited for the Fin acquisition because it's going to be bringing in a service use case to a part of the market that we're not serving today. And so it's going to be very complementary to our current service use cases that we have inside of Agentforce. And so service is by far where we see the most top of action. But what's really -- what I've seen over and over again, especially in our retail space, is that people are looking for super agents. So they're looking for a single agent that they can drive an Agentic commerce experience.

And then if a customer needs to do a return or claim there's something wrong with their product and get that fixed, they want to have it in a single user interface. And now we're seeing this concept of super agents of multiple use cases layered together to go and solve a customer's issue. And the Claudeforce announcement is only going to accelerate that because Claudeforce, we're going to be using, as I mentioned in my earlier comments, some of the agent scripts, which is a really technical term, but in essence, it opens up to a developer.

You can plug in Claude into our Agentforce platform to drive faster deployment. So that's what our FTEs use today to really configure and get agents up and running in 30 days. So Claudeforce is going to accelerate it. And because we have a platform with out-of-the-box use cases, you can start to layer those use cases up. I'll give you one customer example. We have a customer -- we actually have a lot of customers that fall in this scenario, where they'll buy a shopping agent from one vendor and then they'll have a service agent from another vendor. And so they've got 2 separate agents that a customer has to go to, to resolve an experience. And the customers that know I go to this agent for this or that agent for that.

They just want to be able to talk to the agent and get it resolved. Our framework and our architecture allows us to have a single agent experience to be able to jump from a shopping experience into a service experience, all contained inside of one agent that helps lead to more top off because now there's multiple ways that they can consume those flex credits that Bill was talking about. So as people are shopping more, they can use more Flex credits for shopping. As they're driving more service actions, they can drive Flex credits to drive those service actions. Ultimately, the customer is seeing more value and their customers are having a much easier experience in working together with that brand.

Okay. We're coming up on the top of the hour. We're not -- so we're unfortunately not going to be able to get to every single question that is in the queue. We will do our best to follow up with those of you if we're not able to get to your question. The last question is coming from Arjun Bhatia with William Blair. How are you handling model choice and model routing? Will Agentforce be served by a preferred model? Do customers have choice to bring their own models?

That's a great question. And I think it's probably a question really on the minds of a lot of our innovation teams today, ultimately, making sure that the right model is used for the right purpose. But I think it starts with when we built Agentforce and we built the architecture of Agentforce initially, we've always sort of allowed our customers to bring their own model. We've allowed them to sort of optimize what each of the prompts that Agentforce sort of powers behind, you have the ability to sort of pick what models you use to complete those tasks. Because in not every case, you need a premium model to sort of perform kind of routine workflow, if you will.

And so we will continue to sort of open up the platform and allow our customers to sort of bring their own in a perspective. Sometimes they've built their own models and they want to use those models as part of their routine workflow. And so we need to sort of open our platform for that kind of usage altogether. But the further part of the question about will Agentforce be served by a preferred model. I don't think there's a one-size-fits-all kind of structure that we think is preferred or not preferred. What we're going to continue to do is as model innovation continues to sort of exist in the world, we'll keep testing the right models for the right purpose in the right domain. And we will make sure that we will always select probably the best initial model, but allow customers to choice to sort of override to provide sort of the best fit for their organization at large.

So yes, we're -- and I think as we sort of continue to sort of get into the place of arbitrating what is the right sort of offering and the right sort of routing solution, that's an area that we're happy to talk more about at Dreamforce in a few weeks' time around how the innovation platform is starting to emerge.

Here's one thing that I'd add to that. We're working with all the model providers. We're working with open source models as well. And ultimately, what we're trying to do for our customers is we're trying to abstract that complexity from them because all the customer cares about is what's the end outcome that they're getting. Is there a service case being resolved? Is their lead getting generated? Are we providing the right pre-summary brief or field service to the end user so they understand the work that's going to be provided. And our platform allows us to abstract that.

And so it's not the preferred model. It's what's the right model that we need to provide to solve that. And if it's a lower cost model, great. We do a lower cost model, and we can have that value exchange with our customers. But ultimately, our mindset is going to be what's the right model to solve the problem that the customer has so we can deliver the right outcome. And that's where we're grounded right now. And the flexibility of our platform is going to enable us to be able to drive that on behalf of our customers, which is going to be an advantage in the marketplace today.

Yes. I'm going to sneak in one other question that an investor had on the gross margins of all this because I think it's very tied to this question. And obviously, token pricing has shifted a lot. You got a new model launch. The older models might get less expensive or more expensive if they're trying to force you up to that new model. There's a lot of confusion in that, to your point, Connor, on, okay, what do I do from a customer? Which model is right for my use case? We take that ownership on. We help you decide what makes sense for you. If you're really, really set on this is the model I want to use, great, we'll do that. But if we think about the efficiency accuracy curve, which is really what everyone is trying to solve for, we want the most accurate and most efficient outcome and choosing the right model for the right task enables us to take advantage on a gross margin side of the efficiency and token pricing changes that are happening.

So we're being really thoughtful from innovation to pricing and how we're thinking about the underlying gross margin impact of the business on where does that fit? How do we get the best kind of agent routing to the right model to be able to solve for the most accurate, best outcome for the customer while solving on the back end for the gross margin side. So it's a constant moving target, but the teams are working really well together to make sure that happens.

Before we close today, one quick reminder, mark your calendars for our Investor Day is coming up on September 16 at Dreamforce. You're going to be able to tune in for a deeper look at our latest innovation and financial framework. I want to make sure I'm thanking Connor and Bill and Valmik for being here to speak today and overcoming each and every technical hurdle very seamlessly that was encountered. And to our audience, thank you so much for joining us, and we'll see you next time.

Thank you.

Thank you.

Okay. I think we're going to kick this off. We're live. Once again, I'm Brad Zelnick, software equity research here at Deutsche Bank on day 2 of our tech conference in sunny Dana Point, California. Really delighted to be kicking off this session with Salesforce on such an amazing moment in time where we are delighted to have Mike Spencer, Deputy CFO, Head of Finance. Mike, thanks so much for being here.

Yes. Thank you for having me. It's great to be here.

It's -- anybody who's paying attention to the tape, I think Salesforce has really been very prominent, and the stock has had a nice reaction to the news that came out yesterday. Why don't we maybe start there for those that might have missed it, which I don't think there are many. Can you just share the highlights from the strong results that you put up yesterday, what would you focus people on? And very importantly, how does the shape of your acceleration journey look today versus when we first started talking about this a year ago.

Yes. So I made a comment to Brad in the back of the room. I said it's much more fun to come to these when the stock is having a day like it is today. So yes, so -- let me first start with, kind of, the retrospective of last year when we made comments at Investor Day to where we are now. And I would say we've been hyper-focused on driving the execution of the business with the building blocks that we outlined last year at Investor Day. And so what you're seeing now, what we reiterated yesterday in the guidance for H2 re-acceleration as a result of, call it, the last 3 quarters of really, really hard execution, I would say, I'm trying to drive our overall strategy.

So we're super pleased with the results yesterday. I think about it, kind of, in 2 separate buckets, when you look at the results yesterday: First, within the quarter, obviously, we were very pleased to be across the -- all the external metrics. But the metric that sticks up most, at least from my vantage point, is the CRPO and by way of that, the net new AOV, the bookings performance in Q2, which exceeded all our expectations by a long shot.

So -- and that really is a good indicator of the building momentum as we look at H2, look at the guidance that we provided. And more importantly, leads us into FY '28. So the print yesterday, the guidance, the raise in the guidance that we gave, we were very intentional about organic versus inorganic within that guide, which I'm sure we'll get into, but all of that is a result of really doubling down and hitting our commitment that we made around re-acceleration in H2 on the organic business even with M&A as a strategic lever for us.

You were always very popular, but even more popular today. So I thank you even more so.

It ebbs and flows. We'll take it in a moment.

It's really good to see you and good to have you here. And the other really exciting update was this announcement yesterday and the partnerships between Salesforce and Anthropic, which I think there was already some relationship there, but now it's culminated in Claudeforce, which I have a little bit of a tough time. It doesn't yet roll off my tongue. I'm sure it will. Can you just talk about what that offers customers and how it might differ from the relationship and what customers are doing with Salesforce and Anthropic separately today?

Definitely. We -- I had someone last night ask, who brought who to the dance with the relationship. Let me first say just because from a level-setting standpoint, it's really important to understand, our investment in the venture side of our business with Anthropic is mutually exclusive from any partnership that we've got on the commercial side. I think it's really important to ground-setting structure. We -- so the way to think about Claudeforce is, think of it as it's really been an organic process of partnership with them.

For those that use Slack in the room, and if you have an opportunity to use Slackbot. Slackbot is motored by -- or the engine behind Slackbot is Anthropic. And so we've been working with them for quite a while now in a very growing level of intensity is what I would say because Slackbot adoption has, kind of, started to skyrocket quite a bit actually within the user base. And so we've been deepening our relationship over months and months and months.

And what that's led to is a lot of experimentation. We rolled out Anthropic -- or Claude, I should say, within our R&D group over the last 6 months to really experiment and see what that could do to our product road map. And so there's lots of tentacles, if you will, between. And that led ultimately to this idea of Claudeforce where we can make. And for those who use Claude, you are probably familiar with the structure, but there's a set of connectors and whatnot in there and then there's skills that you can prebuild.

And so what that's led to is a, kind of, think of it as an out-of-the-box capability for Salesforce users who decide that they want to do more with Claude and Salesforce that allow you to really plug and play whether that's connecting the assets within Salesforce, Slack, selling our CRM flagship product, et cetera, as well as out-of-the-box skills of Salesforce, think like a salesperson example of different skills that you might use when you're interacting in CRM.

And so we're super excited about it. We wouldn't be doing it if we weren't hearing it from customers. So the genesis of all of this ideation that we've been doing with Anthropic is a result of customers. We do tons and tons of CABs, or customer advisory boards, with CIOs and execs of our customers. And so this has been a, call it, a circular feedback loop that we've been generating with customers that have led to this moment.

It's really exciting stuff. It's good. I know a lot of people are focused on it. When I think of Salesforce, I think like a lot of people, we often first think about core CRM, Sales, Service, Marketing, et cetera. But the company is now one of the largest data and infrastructure software vendors out there in the market. How does the breadth of offering help set Salesforce up for success in the AI era?

So we've morphed our message over the last couple of years around Customer 360 and then we've now grounded with Informatica into the mix. We've grounded that with a dynamic called Data Foundations. Every iteration of our strategy with customers really is about creating, what we call, Layered Cake internally. In that Layered Cake, you've got, obviously, the external model companies as a foundational element, but very closely tied to that is the Data Foundations. It's what we call internally the Data Foundations.

It's made up of Informatica, our Data Cloud, or Data 360, product offering, MuleSoft, which provides connectivity and then Tableau as the analytics engine that sits on top of it. And since the acquisition of Informatica just about a year ago, it's really been, I would call it, a fueling engine to driving and helping customers mature their overall AI strategy, which also, by the way, is leading to some of the developments you're seeing with Claudeforce.

So we feel really, really confident is what I would say right now on the pieces coming together and filling out the portfolio, both from a data foundation standpoint, all the way up through the app layer or the end-use scenario, whether it's Claudeforce or our own applications in customers being -- meeting the customers where they're at. You'll hear us use that terminology a lot.

But as you think about the evolution of our stack, that's really at the core of it. It is helping customers achieve what they want and how they want to work and meeting them at that particular moment. That, by the way, cuts over to our pricing and contract structure as well.

We're going to get to that, but before we do, I think over the last year, we've seen a fair amount of organizational change, leadership change at Salesforce. And I know there's a life cycle to everything. So that shouldn't surprise anybody too much for a company of such nature and significance and heritage. But I'd just be curious, any perspective that you could share on that front. Just understanding how those changes enable the next phase of Salesforce's growth.

Yes. It's -- the changes -- I can even talk about the recent changes, there was changes within the last, call it, 6 to 8 months. It may -- I don't want to be dismissive of it. So don't think this as being dismissive of it. But every change that's happened has been intentional not for the reason of Marc trying to manage people out, but it's been intentional in the concept of folks retire, folks are ready for change. We want to tap folks to take on bigger opportunities. And so every single change that we've made has had rationale behind it that's aligned to the business.

The most recent change, the one I would probably highlight most is Miguel and Alexa, our CRO, Alexa Vignone. And Alexa is a rockstar on the sales side by far and away, probably our leading sales rep, really respected leader and so she was hungry for more and ready for more and Marc really wanted to enable her with that. And then we also want -- she still reports to Miguel.

We also want to leverage Miguel more because Miguel has turned into a force of nature inside Salesforce as well. The sales engine is really, really humming right now. And so as part of that, we wanted Miguel to help us tackle, kind of, the next big problem that we've got. And the next big problem or next big area, I should say, of opportunity is really around consumption and adoption within our customer base.

It's been a focus area for us, but now you're going to start to really see us pour the gas on the fire with that. And Miguel now has taken on professional services, customer support and our -- what we call our Builder Motion. So think FDEs in landscape speak. And so he's really over the next, you're going to see over the next 12 to 18 months, a really, really intense focus from Miguel on that particular aspect as well as leading the overall sales organization to drive that. The rest of the business on the product side is really where there's been more change.

Two big things there. Steve Fisher retired earlier this year. He's been with Salesforce for, gosh, I don't know, 30 years, Marc and him went to high school together.

Left and came back.

Left and came back. And he'd been wanting to retire for a few years and Marc kept talking him out of it and talking him out of it and he was like, "I'm just done." He's got a new grandchild at home, so he was ready to go. And we just hired Rohan Kumar to come in and take over the platform side of the business from Microsoft. I worked with Rohan at Microsoft when I was there. He's going to be a rockstar on the platform side of things. And then we moved Patrick Stokes over. For those that's been able to see Patrick Stokes in the past -- he does a lot of product demos and presentations that are big events.

The guy is amazing, knows our technology, knows our stack, upside down and left and right. And so you'll get to see him more and more as well over the coming days. So the combination of Rohan and Patrick on the app side, Rohan on the platform side, we think it's really going to be powerful in advancing the product stack.

Patrick crushes, he's great and looking forward to seeing all these new leaders in their new roles, especially out of Dreamforce in a few weeks.

By the way, you all should come to Dreamforce. If you haven't gotten a personal invite, here's your personal invite.

Awesome. I know I'll be there. Mike, just maybe on a different topic, enterprises of all shapes and sizes, various industries trying to chase this opportunity in AI. We've seen the token maxing and the spending and customers blowing through budgets prematurely. I would love to get your perspective on investor concerns that AI could potentially be crowding out other spending? Because I mean you guys are an AI company, but you're also an existing software company that's very well-deployed. How did the balance of that all play out to be either a net headwind, tailwind, or neutral for sales?

Yes. I think over the past -- I'm going to answer it 2 ways. I think over the past, I would say, 12 months, let's call it, give or take a little bit. It's been more of a neutral dynamic than it has anything else. I think what you're seeing in a lot of enterprise customer base -- enterprise customers, ourselves included, by the way, customers are re-prioritizing spend to be able to advance the experimentation of the use of the LLMs. On our side of the house, we roughly about 6 months ago, we unleashed Claude in our R&D cycle. It's part of the reason we didn't raise margin guidance on the year is because we're covering some of the token spend that we've got going. And the goal of that really was to, let's see what we could break.

Let's see what kind of advancement our R&D teams can make on accelerating the product road map. Worst-case scenario, we would pull back. Now we're actually in a zone where we are seeing a huge advancement in productivity, but now we're going into refinement mode. So we're going into the zone of prescription model choice for task at hand. And what that really means is you don't need to use the latest and greatest model for every single task you might want to do.

Of course, it's applicable in some use cases. But for the large majority, you're totally fine with the second or third generation model that might be out there. That also includes, by the way, optimization across different vendors. So we -- Internally, we've got OpenAI. We've got Cursor, we've got Claude. So we've got a bunch of different model generators. We're starting to experiment with Grok. All of them have different cost structure. And I think we're -- we are, I think, a good representation of what we see in our customer base. There's a lot of experimentation in our customer base. And customers are in the mindset right now where 2 things are happening. They're experimenting like I just described that we're doing at Salesforce.

They're also advancing or maturing the overall AI strategy, which is why we believe we're seeing -- starting to see the pickup in bookings and why we think Claudeforce or the concept of Claudeforce is really going to catch momentum because of that dynamic of Headless. Customers are starting to understand, hey, like there might not be one UI across my entire user base but they all want to work under, someone use Claude, someone actually want to use traditional app, et cetera. And so over time, we're pretty optimistic, I would say, at this point, that it's going to be a tailwind for overall business, but it's a building -- I think it's a building cadence from here. And over the last 12 to 15 months, it's been more of a neutral dynamic is what I would say.

Makes sense. Like you've demonstrated very nicely for us, the growth and profitability don't need to be mutually exclusive. As we look ahead, what are the largest levers that support further margin expansion? And how should investors think about the trade-off between AI investment and profitability?

So you'll continue to see us be super aggressive from an AI investment standpoint is the first thing I would tell you whether that's internally investing in like the R&D example I just gave, whether that's through inorganic investment, whether it's talent or tech within that bucket, but you're going to continue to see us be really, really aggressive. And that's really a statement about how fast the technology landscape is advancing, and we want to make sure that we've got bets and that we are very diversified from a portfolio standpoint to go after that.

I would say it this way, when you look at our P&L and the contract of our P&L, we've got FY '30 framework guidance out there that says Rule of 50 by FY '30. Within that construct, you've got to obviously believe top line, which I'm sure we're going to talk a little bit more about. And then when you get on the cost side of the equation, I think the way I think about it is really in 2 buckets.

Let's talk about gross margin line first. Within the gross margin line, we expect the cost around supplying of AI within our product set, which is already happening today, we expect to be able to stay relatively neutral on that. Some of that is through incremental premium monetization to cover some of that AI structure, some of it's new SKUs coming through. Some of it is just what I called out earlier, deterministic workflow, not having to pay an LLM if we don't need to, to avoid the cost and/or model selection, us getting into the zone ultimately, and you can see where the train goes of us making the selection for customers on which model they're going to actually use for whatever scenario that's coming up. And so we think we've got a good formula to help control gross margins and make sure that we stay neutral and/or better than where we are today. I'm not saying it's a straight line, you could see ebbs and flows like you saw this quarter versus last quarter, if you're comparing quarter-over-quarter.

But we feel like we've got the right building blocks in place. Then as you move down through the P&L, obviously, no secret to anyone in here. The next really big lever for us is going to be sales and marketing and how do we get more efficient on sales and marketing. Alexa and Miguel are very, very focused on it. We've made a little bit of incremental progress. If you just follow it as a percent of revenue over the past 12 months, and I would expect that trend to continue as we move forward.

The really key aspect of that is as we ramp our AI offerings being able to hit escape velocity from a sales standpoint and increase the mix of self-serve customers, whether that's what we call customers refilling the tank, whether it's consumption-based and/or self-serve customers coming through the website of those various scenarios. We do think there's a lot of opportunity to reduce the cost of sale, if you will. Cost of acquisition of customers through the sales and marketing structure moving forward.

Very helpful to ground us in all of that. I want to talk about Slack if you could. If I reflect back the time of the acquisition, I mean, the investor perception of Slack has really done a, I won't say a 360, I think 180 is more appropriate.

I think Marc said on the call yesterday, the best acquisition we've ever done. I don't know if I'm going to use those words, but like he was very bullish on it.

But it's certainly where we are today, seems very -- he was very prescient at the time. And I won't say lucky because I'll give him all the credit in the world. But it's now evolving to be a key engagement layer for Agentforce; Slackbot is really important. You gave some stats that you can remind us with just in terms of what it's done, I think, with new bundles or pricing. Maybe I've got that wrong. I'm juggling a lot of things and a lot of things that came out last night. But I would just love to hear from you as agents become more embedded in the day-to-day workflows, how important it is as the employee-facing interface? And can it be a more meaningful driver of Agentforce adoption and monetization over time.

Yes. It's -- I think if you fast forward 5 years from now, 7 years from now, Slack could be a super interesting business case in school is what I think you could end up happening. And for us internally, it's been a journey. When I first got to Salesforce almost 5 years ago in speaking with all of you, it was brutal. It was -- sentiment around Slack was as negative as it could possibly be. Fast forward to where we are today and why we leaned into it so much yesterday on the call. The momentum around Slack, as you were just highlighting, is off the charts right now, and it's happening in a couple of different ways. One, engagement is starting to open up a bit more. We're seeing a lot more customers come to the table being willing to experiment with Slack, even if they're a Teams shop.

Obviously, there's antitrust pressure on Microsoft and they're doing some breakup, especially in Europe, even stand-alone sales of Teams. So that's certainly helping. But the experimentation coming from customer base, especially in the enterprise space of willing and wanting to try Slack has really started to open a lot of doors even if they already have Teams in the ecosystem.

And then second, from an advancement standpoint, on showing the art of the possible is how we talk about it internally, but the power of AI can bring to the fold. Within the Slack ecosystem, Slackbot and Headless have really lit up Slack. So Slackbot is our AI engine within Slack. For those who don't use Slack, think of it as a sidecar. So there's an icon at the top of Slack. I could be in Slack Chatting. I click on the Slackbot icon, and then it presents a chat window right next to your Slack -- within your Slack app but right next to your chat window.

And you can do anything in Slackbot that you could do on Anthropic today. It's not Cowork. So I don't want to make it sound like Cowork. But anything you want to do from a Claude chat standpoint or OpenAI, ChatGPT standpoint, you can do in Slackbot. It's powered by Anthropic. And it is super, super powerful, whether you want to pull up conversations that maybe you can't remember which thread it was in and you're like, help me find the last conversation I had with Brad and what we were talking about or a specific topic or something as simple as a very common use case that we use it for and I use it for is something like customer preps.

I'm going to meet with XYZ customer. Tell me everything I need to know about -- I met with BNY Mellon last week. Tell me everything I need to know about BNY Mellon. What's the current conversation, what's the pipeline around them, the execs I'm meeting with, give me the full color, and it will spit back within seconds, a full download and prep document for me heading into that meeting. I don't have to go search for anything. I don't have to go crawl into CRM or elsewhere. So it really brings the power of Slack and then the ecosystem of apps around Slack into it.

That's leading to from a metric standpoint, as you highlighted, Slack being a really big engine to our bookings momentum that you saw. In Q2 alone, Slack was a huge portion of what was a record quarter for us from a net new AOV standpoint and Slack was a huge contributor. If you look at the revenue growth on Slack, which flows through our platform line in our P&L or I guess in our new construct in apps, it flows through, but in the old constructs, it flows to the platform line, the -- it's been growing strong double-digit growth, and we're probably on 5 or 6 quarters in a row now. I mean, I may not have that exactly right. So it is quickly becoming a very, very meaningful growth engine for us for a company.

Well, that gets us maybe now to the part of the matter, which -- Agentforce, a lot of excitement, a lot of anticipation around Agentforce. And it's great to see the momentum I think the stat was what? A $1.5 billion ARR. And if we combine that with Agent -- with Data 360, you're almost $4 billion in ARR as well, $3.9 billion if I have that right?

Yes.

What have been the biggest unlocks from a product and distribution standpoint that have enabled customers to accelerate production deployments and spending?

Yes. The interesting thing about Agentforce and kind of where we're at in the life cycle is that we've taken a crawl-walk-run approach to driving adoption within the customer base. What I mean by that is historically, Salesforce and the go-to-market engine has always been about ACV. Go sell the dollar to the customer and then 9 times out of 10, that AE would then move on to the next sale that you could possibly make instead of driving adoption within the customer base. About 2 years ago, we started to change that.

And in the comp plans in FY '27, our current fiscal year, we actually made it part of the comp plan in a small way, but we had to get water running through the pipes to figure out how we could change the behavior of our account executives. I want to do the exact same journey when I was at Microsoft, and we changed the behavior of the Microsoft force for Azure at the time to what you're going to see next year and with the change on Miguel that I mentioned earlier. FY '28, you're going to see, I think, a material step change in the behavior of our account executives and our account managers in driving adoption within the customer base, which really becomes the engine for Agentforce.

We are seeing an accelerating clip of customers moving from pilot into production. For customers that have already moved into production, the question, if I correlate it back to a metric that we give, which is work units.

Refilling the tanks.

Refilling the tank, but intensive work units have started to run escape velocity. Within that is customers who have moved into full production, we're seeing them refill the tank at a faster clip, 50% of bookings in the past couple of quarters have been from customers refilling the tank.

AOV on those customers is escalating quite rapidly. I think the stat is roughly 2x the growth rate of those customers versus the traditional customers. So it's proven out, even though it's still a smaller cohort of the overall customers that are in the Agentforce that have fully moved into production and are fully ramped, but you're going to see a huge focus from us going forward because that's obviously how we're going to accelerate revenue into the framework.

We're all excited for the acceleration to come. The way we've got guidance set up, I think it exists. It's modest our expectations and I'm always hoping that you overexceed them. But if we look out even further, you've got ambitious plans for an 11% compounded growth rate on revenue through fiscal '30, which takes you to $63 billion. What supports your confidence in these ambitious targets at a time when you're at significant scale in the world changing pretty fast?

Yes. Yes. Let me start with H2, and then I'll parlay that into '28 and beyond. When you look at H2, and this came up at one of our investor meetings earlier today, and it's a very fair question. When you look at our H2 re-acceleration that we're hitting in the second half of the year, we've pivoted that number every which way to Sunday. So we -- in the spirit of keeping ourselves holding ourselves accountable, we -- if you can name an exclusion, we looked at the math, excluding licenses, excluding -- fully excluding Informatica, taking obviously out Contentful and Fin, taking out even when we have headwinds in the business where you got Tableau and Mule and other things. I mean you name it, we looked at it, and we're accelerating in every single scenario. So we feel really, really confident in the numbers.

Most importantly to me within that, is that as you look out at Q4, so it will step up in Q3, step up again in Q4. And you look out to Q4, and you look at the contributors out in Q4 to the accelerating growth that we are seeing, you look at the core, so core organic is growing despite the headwinds from the license volatility we're seeing. So if you said a different way, if I invert that, if you were to look at just the core recurring organic business, excluding licenses, it's actually performing better than when you include licenses. So licenses are weighing things down right now. They're a headwind.

You then layer on Informatica. Once we lap Informatica in Q4, which would be around November 15. It is a tailwind overall growth for the company. We've been -- we're beating on Informatica both on top line and on accretion, and we've been really successful in pouring gas on the go-to-market engine on Informatica. Then you can layer on the additional acquisitions. I'll leave those out for a moment. What we're really excited about, though, is the consumptive -- and the core of your question is the consumptive nature heading into FY '28. So we're at the very, I would say, early stages of whether it's Headless, whether it's Agentforce, whether it's some of our other consumption-based tools. We do feel like we're in the very early stages of those actually becoming a meaningful tailwind to overall growth.

And so the question really, if you flip it around, you say, okay, you're at, call it, 7%, 8%, 9% today in H2. How do you get to 10%, 11%, 12% next year or the year after, et cetera. And that's really how I think about the equation. So license volatility will calm down. We'll start lapping, especially some of the comps that we've got going right now on license volatility, which will help and then you start to see consumption, the spirit of customers refilling the tank accelerate as we get more and more customers into production. And then all of a sudden, you've got the building blocks to getting us to $63 billion and beyond and some of the newer acquisitions, obviously, will help achieve above that.

Awesome. I mean I think we all love ambitious targets. And to your words, like to see accountability and holding yourselves to account at the same time, if I zoom out and I think of the Salesforce journey from 27 years ago when Marc started the company. How do we ensure that these targets -- I mean we're in the midst of this amazing paradigm shift, how do we ensure that those targets are not constraining you to be as successful and relevant for the next 27 years?

Yes. It's a totally fair question. And the parallel I'll give you on that is it's a little bit of our inorganic strategy right now. Part of our inorganic strategy is our AI investment thesis is that we've got to place a number of bets not only organically but inorganically as well to make sure that we're keeping pace on the landscape to your question. That includes talent. That includes tech. It includes established products within the customer base that are seeing adoption. Fin is probably the poster child of this right now. Fin, for a well-established service cloud is, Fin is an amazing product. They've got a very loyal customer base. If you haven't seen the demos of it, you can go on the website and look at it, the Fin technology and their agent and the capabilities of their agent is amazing for a small company as they are.

So you're going to see us. The reason I'd give you that color is you're going to see us continue to make bets across both organically, you're going to see AI infused across all of our products sets. You're going to see us continue to bring on inorganic bets. And we think we've got -- especially with Headless in the mix now, we've got the right combination and I think to fuel us into the future.

The one thing I would tell you, and we were talking a little bit about this earlier is Marc is a force of nature. And he is spending an inordinate amount of time right now, really, really studying the landscape to ensure that we don't miss something. And he's got him and his product leadership team, really, really focused on that. Marc is constantly looking at companies to learn about technologies, look at partnerships, look at targets, look how we can influence the internal product road map. It's also why I said earlier, we adopted Anthropic internally in our R&D group is because we're really focused on accelerating the product road map to make sure that we keep pace. So I would say, whereas optimistic as we can be about being able to keep pace with the market right now.

It's a fair way to frame it. I think Marc gets credit for a lot of things and being on the forefront. I mean, he was the first one to really talk about agents, frankly, in this AI journey and many more to come. I think you also -- you talked about Slack, we'll give them a lot of credit for identifying assets out there that make a lot of sense. We were familiar with Fin pre-acquisition and very impressive customer success, underlying technology, a lot of good things happening there. So maybe if we can pull on that thread a little bit. How do you see Fin expanding Salesforce's addressable market? Where do you see the greatest opportunity for cross-sell monetization across Service Cloud, Data Cloud, Agentforce, the entire portfolio?

Yes. Fin is a unique asset in that it is super complementary for Service Cloud. Think of Service Cloud as the enterprise, kind of, heavier construct of driving agentic structure and think of Fin as, kind of, the lightweight out of the box, plug in and go for the lower end of the market. I think the thing -- there's 2 things that are going to happen. First, on FIN itself, we've created a structure inside of R&D organization called Salesforce Labs and the goal of Salesforce Labs. We kept a guy by the name of Aman, blanking on Aman's last name now. Alexa, keep me honest, but he was the CEO of Regrello. He's come on. He's running Salesforce Labs for us. And he is -- the goal of that is to incubate the businesses. So within Salesforce Labs, at least right now, is going to be Fin, Qualified and Regrello. And the goal is to really incubate them and not crush them, if you will, with the way to Salesforce to make sure that we get those businesses running down the path of the M&A plan that we put together to justify the acquisition.

Then the second phase of it is going to be -- and by the way, that includes go-to-market magic that we tend to do with all acquisitions. Take that one step further. The next phase of that is really going to be taking Fin and then infusing Fin into Service Cloud structure as well. There certainly is opportunity to help Service Cloud within that, and how do we start to help Fin move up the stack.

The example I love to give folks is one of Fin's biggest customers is actually Anthropic. And so Anthropic is in a massive growth phase. And so at some point, we're going to have to figure out how to help them mature through the life cycle of becoming a larger tech company within because they're not going to want to move off of Fin. And so they're already active work stream on that, trying to figure out what that path looks like. And you can imagine a hybrid world and/or a migration path into Service Cloud or graduation path into Service Cloud full steam. So that's, kind of, how we look at it. We do think it's very, very complementary because Service Cloud in an SMB space is not super strong right now. So we do feel like, at least near term, you're going to see a lot of synergy.

Awesome. Maybe thinking, Mike, about getting all these great technologies and capabilities, delivering the last mile to the customer and specifically, as we think Agentforce, and I know it's not just Agentforce, but as adoption expands, how important is the forward deployed engineering motion and partner ecosystem and helping customers bridge the gap between AI experimentation all the way through to enterprise-wide deployment? And how far along are you in building out the FDE capacity?

So I'll share a funny story with you all. It was probably -- gosh, it was probably a year ago now, I don't even remember. And I was flying. There was a number of us flying with Marc, and Marc has obviously got his own plane. So we're flying with Marc and we, for 4 hours on the plane with Marc flying from -- I think we were going to New York, San Francisco or something like that. It sounds like it's fun, but it's got into perks, but it could be challenging. But the reason I share that story is because he's got 2 TVs on his plane. And he literally on full steam or full blast on both TVs for 4 hours had Palantir, Alex Karp, preaching about FDEs on stage of various -- across various different presentations that he did for 4 straight hours.

And I've never gotten more education in such a compacted amount of time on FDEs. And I share that because that's a little bit of a lens into Marc's mindset right now on how critical and important driving customer adoption is especially in a world where customers need the help, frankly, that's the feedback we get constantly from customers is there just so much tech that they're trying to swallow and digest right now that they need help. They don't know where to start. They don't know where to go, et cetera. We're going through it, by the way, with our own adoption of Claude and OpenAI internally where we're trying to roll out Claude and like I have Claude Cowork, most of my organization does not have it yet. But even then, I don't -- would I consider myself an expert prompter? No. Would I consider the different use cases? No. So there's a lot of education that needs to occur. And for us, and why we made the change in Miguel, really getting embedded with the customers.

Think of it almost as, kind of, the old SI consulting model, where you actually put a team in the customer and you're sitting side-by-side with the customer, helping them solve the problems get it launched, et cetera and handhold them through the process. And so you're going to see right now internally, we've actually got a bunch of different flavors of -- we call it builders. So FDEs, builders, it's a synonymous term. But we've got a number of different flavors. We've got solution engineers. We actually have the formal term of FDEs. We've got -- we've got our professional services organization, then we actually have developers inside our customer success organization. So we actually have flavors of different FDEs across the ecosystem, and we're actually working really hard to consolidate them and get them into one motion and how we invest in customers.

Now we're actually starting to look at it more as this investment -- an investment as part of larger contracts that customers are doing. So you're going to see us really contemplate trade-offs of, hey, can we put a few more FDE resources on the customer in exchange for a bigger Salesforce commit or a customer making a bigger commitment to Salesforce, I should say, over time because the payoff of accelerating adoption and consumption is going to far outweigh us trying to monetize a few people for a couple of hundred thousand dollars near term.

Awesome. Switching to a different topic. You touched on this a little bit earlier, but I want to dig deeper just into the way that customers buy and pay Salesforce. How should investors think about the evolution from subscription paid in advance to consumption, maybe AWUs, agentic work units, that you talked about -- and ultimately, what Fin was doing, outcome-based pricing. Like how do we think about that continuum? And what needs to happen for those models become material contributors to the overall Salesforce.

So as a finance guy and owning FP&A, I've got product finance under me. I can tell you, it is an anxiety filled architecture right now of different pricing structures and contract frameworks. We've been very, as most of you know, very intentional about experimentation with customers. A lot of that is fed from our different advisory boards that we operate with customers and experimenting on what works well and what doesn't as customers, kind of, go on their own journey of how they want to consume AI. For us, I think about it in 2 buckets. There is the actual pricing structure and then there's the contract structure.

On the contract structure side of the equation, we have a couple of flavors happening right now. We've got our traditional way where customers will make a commitment, they'll pay us annually, et cetera, works well.

The second one is what we call AELAs. You heard Miguel talk about that last year, I think at Investor Day. Think of that as the All-You-Can-Eat contract for customers where they sign on, they're like, hey, I want Sales Cloud, Slack, et cetera, et cetera. Great. We'll let you, kind of, go over to an all-you-can-eat buffet. The latest one that we've actually been trialing and you're going to see us start to lean into it a bit more is what we call Salesforce Commit. Salesforce Commit is more of the traditional hyperscaler model where customers can say, "Hey, I'm going to spend $10 million over 3 years." And then they can consume it as they go, they can buy seats. They can buy flex credits, et cetera. They can be consumption, it can be seat-based.

But it's customers getting a discount for the overall spend, but not necessarily having to have it all mapped out on day 1. It's very much how the hyperscalers operate, so customers are very accustomed to it, but it's a new muscle for Salesforce.

Then I switch to the pricing side of things. There's a few things in motion. We have traditional seat-based pricing. We've got consumption-based pricing, and you're going to hear that more in the form of flex credits is the way we'll talk about it. And then we are actually experimenting right now with outcome-based pricing. Fin is outcome-based pricing. And so you're going to see us really start to experiment with that more. The key with outcome-based pricing, which many of you are probably familiar with, is you've got to be very objective about the measurements that are driving the outcomes. And that always tends to be the challenge. When you're a smaller start-up like Fin, it's easier to manage. When you get into the broader ecosystem, outcome-based pricing gets a lot more complex.

The way if I compare and contrast, if you look at Sierra. Sierra does outcome-based pricing today, but they actually only give you a couple of different flavors of the outcomes that you can price your contract around. That's how they keep it contained. So you can see us -- you'll see some experimentation around that as we look to expand it. On the consumption side of things, we are moving quickly towards trying to figure out and provide consumption-based structure across a lot of our portfolio, at least where applicable outside of the seats. For Claudeforce that we just launched, it's probably going to be more premium mix. You can get it as part of Agentforce 1 edition or that you can buy an add-on. But you could also buy a full consumption if you want it.

We think it's going to be more limited adoption on a full consumption basis, but you can buy it that way. Over time, most of the revenue today is still ratable. When I say most, 95-plus percent is still ratable today. I think to being in a zone where consumption is a more material portion of overall revenue mix, I think you're looking at probably 3 to 5 years still from now. I don't really anticipate it changing materially in the near term. You'll see more and more consumption come into the fold. But to be an overall material mix of our revenue long term, I think it's probably 3 to 5 years away.

Mike, this has been awesome. It's just about the right time to stop, but -- to end it here. But before I do, in closing, what should we be most excited about heading into Dreamforce?

Yes. I think there's -- I'll speak from a personal standpoint. I think there's a few things. I think you're going to see a lot of really, really great technology at Dreamforce, right? It happens to be the case every year, but I feel like this year, we're, kind of, at an inflection point, especially with Headless coming into the fold. You're going to see some things that I personally would say it will probably blow your mind a little bit on how AI is being infused into the product set. Whether that's Slackbot and the capabilities on Slackbot, whether it's Headless, whether it's some of our newer acquisitions, but we've got a lot of bets that we're really, really excited about, and you're going to see that coming to the fold.

The other thing that I think might be just as powerful is we are being very intentional this year at Dreamforce on bringing in external voices. Marc touched on it on the call yesterday. Dario is going to be there. Sam Altman is going to be there. Jensen is going to be there plus a number of broader tech landscape names, bigger names that you're going to be familiar with.

And part of the goal there is to really get the ecosystem around us speaking just as much as you hear from us, us getting on stage and just evangelizing like how strong we think we're in from a position standpoint, we think can be echoed just as loudly and maybe be more powerful if it's coming from the external landscape. So we're really excited about it.

Again, if you guys -- if you're not making plans or you want to come, just reach out, and we're happy to give you guys a pass and you can come to our Investor Day as well.

If there are hotel rooms left in San Francisco.

I don't -- I can't speak to that. I'll let Anna Le on our team speak to that, but that's out of my pay grade.

Well, listen, again, always great to see you. Thanks so much for being here.

Thanks for having me. Yes, for sure. All right.

At this time, I would like to welcome you to the Salesforce Second Quarter Fiscal 2027 Conference Call. This conference is being recorded. [Operator Instructions]

At this time, I would like to turn the call over to Mark Murphy, Executive Vice President of Global Investor Relations. Sir, you may begin.

Good afternoon, and thanks for joining us today on the fiscal 2027 Second Quarter Results Conference Call. Our press release, SEC filings and a replay of today's call can be found on our website. Joining me on the call today is Marc Benioff, Chair CEO and Co-Founder; and Robin Washington, Chief Operating and Financial Officer. We also have Patrick Stokes, President, Applications and Marketing; and Miguel Milano, President and Chief Operating Officer, joining us for the Q&A portion of the call.

Some of our comments today may contain forward-looking statements that are subject to risks, uncertainties and assumptions, which could change. Should any of these risks materialize or should our assumptions prove to be incorrect, actual company results or outcomes could differ materially from these forward-looking statements. A description of these risks, uncertainties and assumptions and other factors that could affect our financial results or outcomes is included in our SEC filings, including our most recent report on Forms 10-K, 10-Q and any other SEC filings. Except as required by law, we do not undertake any responsibility to update these forward-looking statements.

As a reminder, our commentary today will include non-GAAP measures. Reconciliations between our GAAP and non-GAAP results and guidance can be found in our earnings materials and press release.

Before turning to our prepared remarks, Marc and Dario just announced Claudeforce a few minutes ago live on CNBC, let's roll the clip.

This is really the best of both worlds. This is the #1 AI in the world, anthropic and the #1 CRM, Salesforce, are coming together for the first time in an incredibly powerful way to build a new product called Claudeforce. It's really a first in the industry, Jim. We've never seen anything like it. It's completely exciting. I think that this is the way all enterprise systems are going to run in the future.

We found that this combination product that we built together is the most useful thing in accelerating it. Within Anthropic for a long time, we've been accelerating the research teams within Claude. But this is the first time that we've really been able to incredibly accelerate our go-to-market efforts within Claude. And we want that for all the other enterprises. And we want to -- me and Marc, I want to bring it together to everyone. We've already worked with Salesforce in a number of ways. We're big users of Salesforce. Salesforce is big users of Claude code, of Cowork of other tools. We put products like Claude Tag in Slack already, which is a part of Salesforce. And now this combination is a way to gain something that's it's 1 plus 1 equals 3, something that's bigger than the sum of its parts. My Chief Commercial Officer was just demoing this for me with some of the internal stuff we do within Anthropic just half an hour ago, where he asked what are the biggest accounts that Anthropic is trying to close now. Like give me a list of the biggest accounts. Talk me through the risks of each one. All the data, all the information comes from being managed in Salesforce. But the conversation, the interaction, that comes through Claude. And so you can see how these 2 things can be more than the sum of their parts.

Why this is exciting as it's a perfect complement. Anthropic builds this amazing model. And now the model has this incredible user interface, Cowork. And when you take Cowork and then you're able to put it right on top of Salesforce, it's able to bring the data, the applications, the semantics, the agents themselves and build complete applications, a total user interface to let you get all the value out of Salesforce that's been trapped. So customers have put hundreds of billions of dollars into Salesforce. There is a huge amount of value that can be unleashed through this combination.

And with that, let me hand the call to Marc.

Well, thanks so much. I'm really excited to have the call. And it was great to be on CNBC with Dario announcing Claudeforce. We're going to get into that in more details as well.

And before I get into all of that, I really want to start here at the beginning and talk about the Saaspocalypse. I think you can see from the results, net new AOV growth, it's the strongest in 4 years. Skeptics really said that seats would decline. And also that Agentforce sales and service in Slack, all seats grew year-over-year. Skeptics said, customers would leave, but attrition was near its lowest level ever. Skeptics said pricing power would erode and A1E and A4X bookings more than doubled quarter-over-quarter. And contract linked terms improved across all segments and new business and renewals. An agent agentic use, well, use of the platform, it surged 6x, sixfold via model context protocoled calls and CLA calls. Apps are not [ dime ]. In fact, they're growing. And 9 of the 10 AI companies, like you just heard from Anthropic that's standardized on Salesforce, use Salesforce and Slack. They spend up -- their spend is up 435% year-over-year. Frontier models depend on CRM, they make sure that it works, that it all comes together, they'd not replace them.

Let's now -- let me start from the beginning of my script and actually read my prepared remarks for you, so you can really understand our vision for the future of where Salesforce is now, where we're going, and we cannot wait to get all of you to Dreamforce because we have some incredible things to show you.

Look, we're here in Salesforce Tower in San Francisco. I want to welcome all of you, and I'm thrilled to share the tremendous progress we've made and momentum we're seeing across the business. As everyone can see, we just had 1 of the best quarters ever, and we outperformed on all of our key metrics. And as I just went through in all these key metrics in these key areas, this nonsense of this SaaSpocalypse, I think it's time for it to stop. That's why I'm so excited about Claudeforce as well. It really shows how we're bringing together the world's #1 AI and the #1 CRM. It's really the best of both worlds. We're combining Claude's extraordinary intelligence and reasoning with Salesforce's, trusted data and applications and workflows and business rules and governance. And while Claudeforce -- Claude might have been, for example, for developers, now it's also for knowledge workers. And when you do that, you get something in the world has never seen. This is a dynamic, deeply personalized interface that thinks reasons and act. It's a truly game-changing experience for our customers.

And I'm going to get into the highlights from the quarter. But before I do, I want to just tell you 1 quick story. And I'm going to get into more details on this. I just spent 2 months in Europe. And I just had an exhaustive schedule. I won't go through all the detail, but I've never met with more customers in my life. And I was with 1 of our very largest, probably our largest customer in Switzerland, and I was sitting there with them, and we went through and they spent maybe, I don't know, the they spent more than $1 billion for sure, with Salesforce. They spend about $100 million a year with us. And I'm talking with them, and it's a great meeting, it's 3 hours and then at the end, we're like, no, I want to show you this thing. We haven't shown it to anybody yet and we'll show them Claudeforce, because the head of all of our operating units have been using Claudeforce to run their businesses inside Salesforce this quarter. It's 1 of the reasons why we're getting such great performance. And all of a sudden, literally, I could see the customer's mouth just dropped open because it was astonishing. They were seeing that we were able to reveal this value that had been trapped in their systems, okay, right into this dynamic user interface. And they weren't having to hire all these FTEs and all this stuff to get it going, it building these apps on the fly for them using the foundation that they had spent 2 decades building with Salesforce.

Now I want to get back to the quarter. We had record revenue and EPS, strong margin performance. CRP accelerated. It was up about 14%. Look, we generated $1.1 billion in free cash flow, that was up 81% year-over-year. I'm going to come back to that in just 1 second. And our AI business, just incredible momentum. Agentforce ARR, it hit $1.5 billion. Our ARR for AI and data is about to cross $4 billion, really helps, I guess, that we have 15,000 peer account executives, Miguel. And the number of accounts with agents in production grew 70% over last quarter. We had great wins with Deutsche Telekom. I met with them when I was in Europe. FIFA, met them. NTT Data, I have met with them. Cisco, all expanding their AI investment with us. Great companies, by the way, great visions for the world and really exciting executives leading their companies who have tremendous vision for what they're going to do with their company using this new technology. And we also completed a multibillion dollar agreement contract expansion with the U.S. Army. The Army Human Resource Command expects to drive up to 55 million Agentforce conversations every single month, amazing. The Secretary of the Army took me aside and said, he finished recruiting 4 months early because he was able to use not only the power of all of these human agents, but AI agents as well, that was the Army being all they could be. And I'll tell you, we're expanding in the markets and seeing an incredible success. We launched Agentforce IT service in November, and it already has more than 450 customers, including NYU Langone Hospitals at McAfee, conversions from ServiceNow. I think that's very exciting that we can see that more companies coming into our platform. And as they come into our platform where we expand it to be not just sales, not just service, not just marketing, not just commerce, not just analytics but also IT service and other critical functionality, that all of that can be revealed and now rendered through the Agentforce user interface. The more that's in the platform, the more Claudeforce is going to be able to deliver for it.

Now Agentforce Life Sciences is more than 140 customers, including all top 5 pharmaceutical companies, and Agentforce ops is helping companies like Kellogg's and Dell turn manual back-office work and supply chains into agentic workflows. That's the Regrello platform we bought now more than a year ago, it has the ability to deliver and render for a company, red, yellow, green on their supply chain every single day. For Dell, they have 20,000 suppliers. And why that's important is we see supply chain as 1 of the very first applications. So many companies are actually revising for AI because it gives them the ability to actually look into something and do something with an area that really has not moved forward for in decades. With this strong performance and momentum and the strategic acquisitions, especially Contentful and Fin, we're expecting to close in the coming weeks, and we're raising our fiscal year '27 revenue guidance by about $300 million in constant currency.

Now what's driving all this? Well, it's a fundamental shift in the software industry. AI is unlocking value, as I said, across every part of our platform. AI is able to look into the platform and say, "Hey, I can see this application construct. I can see this user model. I could see the share model. I can put all of this together and deliver value in a totally new way." I've never seen a moment like this in my life, but for our customers, the question isn't whether AI is transformative, it's actually how they get there. And I want to talk to customers every single day. That's why I was in Europe. Most of them have not started their AI transformations yet. If you talk to most CEOs they may say they have a project or they have a failed experiment or they were told to build a model, but it didn't work out for them. That is not AI transformation. We're going to see AI transformation, and you're really going to see AI transformation when you start using Claudeforce. And the reason why I feel so strongly about that is I've seen it here at Salesforce. For our operating unit leaders when we gave them Claudeforce, they came back to us and said, "I built this app. I did that. I did this other thing." That was so exciting. They built their businesses on not just a few core systems. They're not going to just rip out how their company just runs to get to they want intelligence delivered inside the software. They're already using to run their business. And that's what Salesforce, that's what Claudeforce is going to do.

So I showed them Claudeforce. I showed them reasoning across their Salesforce data and workflows. I can show it right inside Cowork. I also show it right inside Slack, 1 click to deploy and their path to AI transformation just opens up in front of them. It's pretty incredible. It's a great experience as a software executive to see the customers' reaction. And this is what we're building with our new enterprise harness.

So this new enterprise harness AI for us, you're going to hear a lot about this at Dreamforce. First, you're going to hear a lot about our data layer. You're going to hear a lot about how we've invested in the federation and integration and harmonization of data, you know that's why we bought MuleSoft. It's why we bought Informatica. It's why we built Data 360. You're going to hear about our application layer, of course, sales, service, marketing, commerce, analytics, as I just said, supply chain or ITSM, all the different components, field service, all the application lever in that application layer, then the agentic layer on top of that. And then finally, the interface layer. These 4 layers, data, apps and semantics, agents and interface. When it's turned on together, that is when you start to see the true enterprise transformation. And I'm confident that when you come to Dreamforce, for your own company since so many of them are already on Salesforce, you're going to see how you are going to be able to transform as well.

Now let me show you what this looks like for AthenaHealth. Every time someone needed a case summary or opportunity insight, they had to ask IT to run a report. So they use Data 360 to bring together all the product usage and account data from across their systems to create a single source of truth. Then with AI force, this new user interface harness that sits on top of our core systems, they've connected all of it to their AI, complete with the context and the permissions that already lived in Salesforce. Now employees across every level can get instant answers from real Salesforce data -- from real Salesforce metadata and all of the sharing models and security models and all of this, all of this is done with 0 data retention. That means no data goes to any model. It all stays in your company. We started engineering that in 2023. We have been perfecting over the last 3 years and we audit it and we are sure that, that data is staying with you without ever opening a browser, the usage and the value they're getting with Salesforce is going way up.

This is the future of enterprise software. And I'll tell you what it's not. This is not the SaaSpocalypse paces. This is not the SaaSpocalypse. As I said earlier, we've been hearing about this for the last 2 quarters, these dire predictions about the end of software and how but the models need everything. But none of them have come true for us, and I don't think -- I don't even understand how they could possibly come true.

Look, models, you have to remember, models are probabilistic systems. They're nondeterministic systems. But our system, those 4 layers that we talked about with data and apps and semantics and agents, those are deterministic systems. But when you put those 2 things together, that's a level of value we've never seen before. We've been told our growth would slow, that net new AOV growth would not continue to grow. And yet it's the strongest it's ever been in 4 years and significantly outpacing AOV growth. And by the way, did you know that Salesforce, Salesforce has higher free cash flow, higher free cash flow than Costco, Coca-Cola, Home Depot, Ford, Disney, Pepsi, oh, we also have higher revenue growth rates than Costco, Coca-Cola, Home Depot, Ford, Disney, Pepsi. By the way, all of them are great Salesforce customers, too. I love all those companies and all those CEOs, but we have higher free cash flow and higher growth rates in them. I just think it's important for speaking directly to the analysts who cover us to actually look and compare our business metrics against those others so you can really see what we do and how we're doing. Our seats were supposed to decline. Instead, Agentforce sales, services Slack, all saw year-over-year growth, and we were told customers would abandon us, but attrition is near its lowest level ever. I know that you know that because you're calling all of our customers asking them. So we're told customers would pay less for our products, but what Miguel will about to tell you is they're upgrading, an A1E and A4X bookings, they more than doubled quarter-over-quarter. We're told contract links would shrink. Well, they improved across all segments, and we're told agents accessing our system would kill apps. Well, actually, usage is exploding, agentic use of our apps through model context protocol calls and CLI calls also surged sixfold, 6x. And we're told the frontier models would eliminate the need for CRM. Instead, they depend on it, and we're showing how 1 plus 1 can equal 3. Nine out of the top 10 AI companies in the world use Salesforce and Slack. You know that. You're talking to them, companies like Anthropic, for example. Well, these companies, these 9 out of 10 top 10 AI companies, they've increased their spend with us by 435% year-over-year. Look at that great company Replit, you saw them in the preshow. It's an amazing company. Their sales team lives here in Slack with Salesforce as their source of truth, they've tripled the number of Agentforce sales seats. They're building custom apps that connect directly to Salesforce data and workflows. It all works together. By the way, Replit is a great tool on Salesforce.

As you can see, AI isn't raised replacing Salesforce. It's unlocking more value across all 4 layers of our platform. again, across the data layer, the apps layer and the semantic layer, the agent layer and the interfaces. Now we all know that if your data is not right, your AI is not right. I've seen that over and over again in customers where all of a sudden, we'll go in and try to show them some magic and all of a sudden, we'll go, "Well, are you really paying attention to your AI, because your data quality doesn't seem to be exactly right." Well, all that data inside Salesforce now provides hundreds of petabytes of critical context so agents can actually understand your business. We're doing more to help customers clean up their data to make sure that it's well harmonized to make sure that it's -- they have the right level of data cleansing, building the right data warehouses, then our apps, well, they can be more valuable than they've ever been because now they don't just run your business, they're running all of your agents as well. Think about it. There's more data insight Salesforce than any human can comprehend. A human works 1 record at a time, an agent works around the cross -- the clock. It sees everything, not 1 thing. It reads across millions of records to find the perfect opportunity, forecast the quarter and close cases.

We know we've been using these models now for almost a decade. For a decade ago, I would sit in my staff meeting, and I would ask "Einstein, tell me how is the forecast going to turn out." And in many cases, it do more than our folks. Why? Because it's able to see everything.

So here's another great example, Uber for Business. They've got 30 onboarding specialists and a mountain of inbound leads nobody was ever going to reach, their rules, their workflows, every deal that's ever gone through them. They've used Salesforce since the start of Uber, and all of that lived inside Salesforce. Now this is how they sell. So they paint -- just pointed Agentforce at the pile, got to work, looked at their whole history of their whole company, all the institutional memory everything that's inside Uber, now 6 weeks to launch within 2 weeks, 60% more leads and converting. This is that power. We're introducing hundreds of packaged agents that understand your business and industry right out of the box, so they can start delivering value within days, not months. And now with AI force, we're bringing all that to any interface, this is the magic moment. People don't have to come to Salesforce to get work done, Salesforce comes to them. And Claude and Slack and ChatGPT and Teams, wherever they want to work, and it's a shift from software as the interface to software, powering every interface.

Of course, this is already happening in Slack, which had a phenomenal quarter. It's the best acquisition we ever made regardless of all the financial analyst reports that I read about 6 years ago. When we bought it, we won't go through the details, it's water that's now all under the bridge. But let me just say this. When you look at Slack today, it is where work happens. And it's where these AI lives and work and could do things. And not only is it where human work happens, now it's where you can build together to with Slack Codes. AI isn't trapped in a single-player chat window. It's a teammate that works across all your people, finding the right agent, the right action, all the flow work, fueled by all the enterprise context that lives in Salesforce, the conversational context in Slack. Companies like Anthropic, OpenAI, Lovable, Replit on and on and on. I think there's more than 1 million companies on Slack now, are all building their agentic products directly into Slack because Slack is where their customers want them. And Slack bot is our fastest adopted AI product ever. Five months after launch, it's got 1 million active users, up 150% quarter-over-quarter. If you don't have Slack bot turned on in your Slack instance and you're not using it every day, you are really missing out. It has the ability to read across all of your Slack data, and you were going to have insights in your business, you just were not able to have before. It just makes you a better executive.

Robin Hood is a great example of what's happening. Robin Hood needed AI adoption across its whole workforce, not just engineers. So what Vlad did was he put AI inside Slack, the place where the entire business comes together and connects to Salesforce. Google, Okta and other systems. And now any employee can use Slack bot to surface past decisions and conversations, solve problems without engineering support and Slack isn't just the best place to work with AI. It's also the best place to build with AI. We just introduced Slack Code, a shared space for teams of agents and humans to build together right where they already work. So when you combine Slack Code with Slack bot, it becomes an incredibly powerful multiplayer IDE.

Okay. Well, as you can see, across our whole platform, we're redefining how our customers build, deploy and scale AI, and we have a very special guest today on our earnings call, and our friend, David Friedberg is here. He's got this great next-generation bioengineering company, Ohalo. Welcome, David. Fantastic. Welcome to our earnings call. Great to have you.

Thanks for having me, Marc.

And they're changing how we feed the planet with a breeding technology that boost crop yields by 50%, maybe 100%. [indiscernible] sent me a can of seeds. I don't know if I'm allowed to talk about what I got or not, but it was pretty awesome. And I'll tell you, David had chosen a CRM product when he first started Ohalo and -- how does it pronounce? Oholo? Ohalo? Mahalo?

Ohalo. I should ask you. It sounds like a Hawaiian word.

It sounds like it, but it's not. It's based on an important actually part of bioengineering history, right?

Yes. Well, there was an agricultural find in the Sea of Galilee where they found this whole pot of seeds 20,000 years. So kind of rewrote what we thought about the history of agriculture and plant breeding. So we named it after that little village.

Fantastic. Well, yes, you're right. It sounds like a Hawaiian word but it's not. And I know that you started with another CRM product, we won't talk about which -- what it is and -- but I remember when you called me and you said, well, maybe I'll try to vibe code this whole thing, whatever. And then a few months later, it turns out your...

Yes, we did a little vibe coding. I did a vibe code to make a CRM product over a weekend. And you quickly realize just how much it takes to do maintenance, to do accounts, to do security, there's just much more to it. And we realized pretty quickly, there are certain, I would say, software applications that are maybe verticalized where you have to build custom workloads that make sense for your particular business, but then there's platform software. And platform software, we realize pretty quickly is really what we need to build our entire company around. And Salesforce is what we kind of decided to go with on CRM, we're already on Slack. We're not going to go vibe code Slack, we're not going to vibe code CRM. But then we build custom workflows, custom interfaces, custom prospecting tools that my sales teams. I've got agents that are out scanning public county databases, trying to find where farmers are planting what crops, what their names are, what their contact info is. I've got agents doing all sorts of prospecting, that interface gives my sales team leads that integrates with Salesforce. They can then go run the standard workflows that are a little bit more standardized across verticals, and we could build all of our custom workflows in parallel and in concert with the Salesforce tool.

Well, it's been very exciting. And I'll tell you, I'm thrilled to have you with now a large Salesforce customer and...

One of your smallest, but yes.

Well, I would say large in terms of presence, in terms of -- everybody knows you now Dave because your podcast is so popular. The #1 podcast in the world is what I heard.

In San Francisco. For now. That's right.

But what I want to know is this, so when you do think about the future and you think about your company and bringing the best of AI and the best of CRM together, how do you plan to do that?

Again, I think it goes back to this point about where do we create value? I'm not going to create value as a business by doing something that everyone else does. If everyone else is using Microsoft Excel, I'm not going to go build Microsoft Excel. If everyone else is using an ERP tool, we integrate with NetSuite, and I know that you guys have an integration with NetSuite. We're not going to go build an ERP tool. And we're not going to go build Slack. We're not going to build communications and messaging. We're not going to build CRM. We're going to build custom interfaces for doing plant breeding. We're going to build custom interfaces for prospecting customers. We're going to build custom interfaces for selecting the right product for a farmer based on my satellite and radar imagery that shows me what that farmers field is going to look like next season, making an estimation and then making the right product selections for him. So the interface to do that for my sales team for my customers, that's where we add value. And then the actual following up with the customer, connecting with the customer record making that transaction record, storing it. That needs to be done safely, securely, it needs to be done with the right account settings, the right access. That's what we're not going to go build. And that's really where the partnership works.

Now when we make a selection, it has to be a selection with standard platforms that we can work with that integrate that have the integration capabilities that have the ability to service and support. I'll give you another example. Just internally, we use Slack, right? That's our primary communication messaging tool. And in Slack, we've got an Ask HR channel now. And so we basically use Claude to set up an Ask HR channel that anyone in my company can ask the HR questions that they would normally spend time calling up the HR people saying, "Hey, I got this question, I got this question." Ask HR has access to all of our internal HR documents, and it knows who you are, whether or not it can access your personal information. It knows standard benefits policies, and I can just answer all your questions for you through Ask HR. I don't need to go build a communications tool to do that. I can leverage Slack and I can leverage Claude to deliver that internally. It's another good example of kind of the sort of thing where we can build something custom using our internal benefits, our internal HR policies, but then we can leverage the platforms and Claude to deliver that.

Okay. Now because I'm a competitive person, I have to ask you this last question, which is kind of where I started. This all started because we were just talking on the phone about something else, we're personal friends. And all of a sudden, you said to me, yes, I'm using the CRM product. And I'm turning it off completely. I'm not going to use this product anymore. And then I was like, well, why? So can you just tell us the story, you had already invested in the CRM platform. This is 1 of our competitors. And then -- but you decided not now, you don't have to tell us who it was, but what are some of the features and functions that you saw in Salesforce that then gave you this ability to achieve all this productivity that you so badly desired?

It wasn't sold as great a sales guy as you are. I wasn't sold installed in our first conversation. I was still pretty reticent. I'm like, Marc wants me to do this. I'm going to do it for Marc. But I'm not sure, okay? This is what I told my team. And there's a guy, a great guy named Ben. Ben worked at my shop and Ben used Claude. He used Cursor, and he spoke to your Salesforce guys, stood it up. And what took us months of going back and forth trying to customize this other CRM tool and build the application and workflows around it that we needed. We were like pulling our hair out we're like, let's just build it all from scratch, and that's when you and I talked. So Ben used Claude and Cursor, got into Salesforce and was able to get everything stood up in under a month. And so I wasn't sold until we did it. And then I was like, okay, I'll give them our credit at this point. And maybe we should send them a check and pay for the product. But I would say that was the...

By way that a huge...

You're a big marketing guy, but the product and the capabilities...

Thank you, David. Appreciate that.

That sold itself actually. Like when we got in there and Ben was able to do this in under a month, and he told me he did the whole thing in under a month. Just Ben. And he ran the whole thing.

And I think that has been the huge shock for us that these new next-gen tools, the Cursors, the Replits, the Claudes of the world, make our product better that all of a sudden, it can customize the data, the metadata, set up the preferences, do the administration, do things that maybe you needed a bunch of service folks to kind of do. Now all of a sudden you're getting going in a month, maybe it might have taken a view before, 6 months to a year, all of a sudden now you're fully automated and your ability to add new functionality and do new things. This has provided this kind of very strong platform. And as I said, it's a terministic platform. And while Cursor is more nondeterministic and probabilistic on the front end, it's helping you to administer it, maybe Bill starting to build new screens for you, analytics, but nothing is going to be better than having this solid foundation.

I was probably early on a SaaSpocalypse train, as you know, because I know you listen to my show. And I always said, like, I think SaaS was this temporary phenomenon between the founding of the Internet and the start of AI. But I think that it's a little more nuanced than that. It's probably this verticalized SaaS where you're trying to standardize a vertical on a bunch of workflows, which actually destroys value in that vertical because everyone is now doing the same thing in the same way. So no one can differentiate in that vertical. We've actually dumped all our vertical software -- verticalized software. We make all of our verticalized SaaS in-house. But the horizontal, the platform, Salesforce, standardizing on Slack. This is really where I think we've realized that's not -- that's actually going to get bolstered and it's more valuable with all the other capabilities we can now build around it. So I'm definitely sold in that sense. I don't think -- I think there's still a SaaSpocalypse with a lower case S rather than the upper case S for verticalized tools, where I think AI really allows you to rebuild something that's unique which creates value for your business in a vertical.

Rebuild it on Salesforce.

Yes. That's right.

Great. Thanks so much David for being with us. We really appreciate you coming in today.

All right. We're so thrilled to have David here, and we're so grateful to have him now as our new customer. And I just thought everyone would love to hear that story because I think as everybody has heard David has been wondering how will he use enterprise software in this new company, where does he achieve that productivity. And I've just never been more excited about our future. The scale of the opportunity in front of us is extraordinary. You're going to see it at Dreamforce. Not only are we going to have great entertainment, probably the better, best entertainment we've ever had. So we can't wait announce that. But also we've got probably the best products we've ever had. And you're going to see it in just a couple of weeks, it's going to be a huge accelerator for Salesforce for all of our customers. We're going to bring 50,000 people here to San Francisco and show them some unbelievable capabilities. And we're going to show you some of the most exciting transformative technology we've ever built. It's the future of enterprise software. And I'll tell you, we're deploying it not just for the Army or for the IRS or for all the major U.S. agencies, all these incredible new customers that Miguel has signed up from Uber to Deutsche Telekom, FIFA you're going to see not only meet all of them, but you're going to see an incredible lineup of speakers at Dreamforce. And yes, Dario is going to come in there. It's going to be in my keynote. You're also going to see Sam there. Jensen is going to be there. Actually, all the major AI leaders is going to be there. David is coming, if he's bringing all his friends. And you're going to see how great companies like Amazon, NVIDIA, F1, L'Oreal are bringing humans, agents, platforms together to connect with customers in new ways that were never possible before. We're looking forward to seeing you all there. And before that, we've got you, Robin.

Thanks, Marc. I appreciate it. Your enthusiasm is in infectious. So I'm going to go with a big speedy here because we want to be sure that we value all of your questions and want to be sure that we can tell you everything we can about the quarter.

First and foremost, we delivered a record second quarter. AI is amplifying the value of our platform. This isn't just the technology shift, as you've heard, it's a reinvention of our customers' work and it is fueling our growth. So I want to take you through a few of the proof points. It starts with our operational focus on net new AOV, which is driving our reacceleration. Exceptional Q2 bookings, along with near record low attrition, drove our Q2 CRPO beat. Half 1 net new AOV growth significantly outpaced AOV growth, keeping us on track for second half organic revenue reacceleration. Contract lift terms for new business and renewals are increasing across all segments. Slack continues to shine. Q2 Slack net new AOV was the fastest quarterly growth since acquisition. AI innovation across our platform is delivering measurable value. Agentforce ARR reached $1.5 billion, as you heard, driven by Slack bot and Headless launches and continued AI momentum. Upgrades to our premium Slack additions have tripled since we launched Slack bot. Our consumption flywheel is turning. AW use continue to scale as customers drive agentic use cases across the platform. In Q2 alone, customers drove 3.2 billion AW use, up 97% quarter-over-quarter. And Slack bot users were up over 150% quarter-over-quarter. With this clear demand, we are making it even easier to adopt. Our premium bundles give our customers access to our innovation in real time. Bookings from Agentforce 1 edition and Agentforce for apps more than doubled quarter-over-quarter. Headless takes this even further, bringing Salesforce to the surfaces where teams already work, Slack, Claudeforce and with AI force, many more to come. MCP and API usage grew significantly in the quarter as customers embed sales force intelligence, the data, metadata and semantics directly into their workflows. You can see the momentum in our results from the quarter.

Q2 revenue came in at $11.35 billion, up 11% year-over-year, above the high end of our guidance in constant currency, driven by continued momentum in Slack and Agentforce. And we continue to accelerate Informatica's business. Subscription and support revenue was $10.82 billion, up 12% year-over-year in nominal and 11% in constant currency. Our top line performance reflects the breadth and resilience of our portfolio. Agentforce apps delivered solid growth, anchored in momentum in Slack and resilience in sales and service with early signs of recovery in marketing.

While it is too soon to call this a sustainable trend, we're very encouraged by the success enterprise customers are having on our next-generation marketing product. Headless platform, Data 360 and other growth was fueled by continued momentum in Informatica and Data 360, partially offset by license revenue headwinds and volatility in integration and analytics. CRPO accelerated quarter-over-quarter, with $33.5 billion, up 14% in constant currency, that's 1 point ahead of our guide, driven by strength in Slack, Agentforce and Data 360.

Our continued focus on operational discipline is evident in our profitability. Q2 non-GAAP operating margin was 34.1% and GAAP operating margin was 20.5%.

As customer zero, we are putting Agentforce to work across our own business. Salesforce's help agent have surpassed 5 million customer conversations with 64% resolved autonomously. Slack bot is driving 8.1 million hours of annualized productivity gains for our employees. That's more than double quarter-over-quarter. We generated $1.1 billion in free cash flow. That's up 81% year-over-year. Execution of our 25 billion accelerated share repurchase continues. We expect to repurchase at least 14% of shares outstanding through this program thus far at an average price of $176 per share. Our responsible capital return strategy underscores our conviction in our future.

So turning to our outlook for the year. I want to walk you through the components of the $300 million constant currency raise to our revenue guide. We are raising our FY '27 revenue guide for organic performance by $100 million. This is driven by the continued momentum in Agentforce, Data 360 and Slack. These areas of strength offset the continued volatility in overall license revenue. Alongside this organic growth, our raise incorporates a $200 million expected contribution from the anticipated closings of content and then in the coming weeks. Of note, our FY '27 revenue now incorporates a $100 million FX headwind compared to our prior guidance. This brings our updated FY '27 guidance for revenue to $46.1 billion to $46.4 billion, resulting in subscription and support growth of slightly above 12% year-over-year and slightly under 12% in constant currency. We're holding our non-GAAP operating margin guidance at approximately 34.3%. And we're updating our GAAP operating margin guidance to approximately 20.1%. We continue to expect operating and free cash flow growth of approximately 4% to 5% year-over-year.

So moving to our Q3 guidance. We expect Q3 revenue of $11.42 billion to $11.5 billion, growth of approximately 11% to 12% in constant currency. Q3 CRPO growth is expected to be approximately 14% year-over-year in constant currency. I do want to note that this does not include incremental contributions from Contentful and Fin, as we await deal closing to ensure accurate balances are reflected.

In closing, building on the momentum from a strong first half, we are delivering record revenue and cash flows this year. We look forward to sharing more on our latest innovation and financial framework during Investor Day at Dreamforce next month and connecting with many of you here in San Francisco.

So before I turn the call back over, I want to thank Mike Spencer for his incredible leadership of Investor Relations. His candor and clarity earn the trust of the investor community and I'm personally grateful for his partnership. Luckily, he's not going far. He's now taking on a broader role in our finance organization. I also want to extend a very warm welcome to Mark Murphy as our new Head of IR. He's covered Salesforce as a top sell-side analyst going all the way back to our IPO. So he knows many of you in our story very well.

Back to you, Marc.

Thank you for the kind words, Robin. With that, we will move to questions. [Operator Instructions] Sophie, please queue up the first question.

[Operator Instructions] We'll take our first question from Kirk Materne from Evercore ISI.

Yes. Thanks very A lot of great things going on this quarter, but I have to ask about Claudeforce. And Marc, I think specifically, I'd love to understand how you and Miguel are going to be taking this product to the field, what are you doing with Anthropic to make sure that this gets out and gets to the hand to your customers as soon as it's available. And I assume we can be seeing Anthropic and Salesforce salespeople working together to bring a lot of value to both your collective customers.

Well, we're really excited about this product. This is the best of both worlds. It's the #1 AI meeting the #1 CRM and putting them together. As I said, when you do that, you get this incredible result, a radically different type of interface on top of Salesforce's entire platform. It's really powered by this incredible new AI harness that you're going to see at Dreamforce AI force. And that harness is able to power not only Claudeforce but a new version of Slack that you're going to be seeing as well as Coworker, which is something inside our lightning interface and other amazing things as well. But once you have that enabled, then you're able to really unleash the power of our data, our applications and semantic layer, agent layer and all of it comes to light in a whole new way. And I really saw it myself, as I said, in Europe, where we would really show these exhaustive demos to customers and when they could see that all this value that they have in their systems that has been trapped becomes untrapped, but you also saw it here at Salesforce. This quarter, we actually released it to many of our executives all over the world. And many of those executives were the ones who are actually demoing it directly to those customers that it was sales executives are my most senior sales executives, my operating unit leaders we're showing Claudeforce to our top customers already. And you're right, we'll also work hand-in-hand with in Anthropic sales executives as well and we'll do whatever we can to make sure that this becomes a huge part of our sales ecosystem. I think this is going to be an incredible opportunity for both companies. And I'd love for Miguel to come in and talk about his vision for how this has come together, and Miguel has been a huge driver of the relationship with Anthropic as well as really broadening the capabilities of both companies.

Thank you, Marc. Listen, I've been obsessed with this surface. I, myself and a number of my leadership team have been using very profusely the surface connected. We've done it manually, connected the SEP servers, created skills on our own. So I just can't believe that we have now this product available. We are rolling it out to the whole enterprise or the whole company. And then we're going to make it path. We're going to make it available to GA in September to everyone at Dreamforce. Every single seller of Salesforce will be demoing this product as 1 of the surfaces. By the way, we love Slack as 1 of the surfaces, but not everybody has Slack. They are complementary. Slack. You are working with your teams is a multiplayer surface with Cowork, you and Claudeforce or Salesforce before Claude. We are simply going deep and research and understanding your business. By the way, you are able to engage also through Slack through e-mail with your team members, it's going to be very powerful. Every customer will be able to buy it. They will have to upgrade to our premium additions and it's going to be a new wave of demand. I'm very excited about it.

And Miguel, just give us just a couple of anecdotes because I just know like -- when we turn this over to several of your European leaders specifically, and obviously, I was in Europe for that time, and they started using it, just even at the most basic level, the kind of things that they were building and the kind of things they could see about their business and how to drive their business forward really changed. Can you explain that?

So currently -- and by the way, this is what I want all our sellers to do. We're going to be living in 3 surfaces, what we built, and this innovation came from Europe more than the U.S. I'm very excited about that. We basically created a cockpit to understand the business.

So you're saying that Europeans can innovate.

Sometimes, we can. Yes. So we build our own agents based on our Claude and that agent expect, they are our deputy CROs for all our businesses, and they inspect the business, the pipeline, they tell us if there is risk in the month or in the quarter. And I do 1 thing that is -- Marc, you would love this. I talk -- by the way, I do this on Slack. This is 1 of the surfaces. I do this in -- I mean, at the time, Cowork now is Claudeforce. But I also do it in an app that I built with Claude Code. So depending on the moment I use 1 of the surfaces. But I do 1 thing that is very powerful. Every time that I visit a country, and don't tell anyone because I don't want my team to know. But essentially, what I do is, okay, I'm visiting Italy. I mean we have an amazing leader there, Vanessa Forteza, she's our [indiscernible]. And I said, okay, give me all the opportunities, open opportunities for the month and then send a Slack to every account executive copying the whole chain all the way to Vanessa, telling them, look at the opportunity record, look at everything, make it sound like it's me. I review all the e-mails all this like before they go. And in [indiscernible] 50 very customized Slack messages to the AEs, asking them if I can help, if we can do anything, it's incredible. The engine is very smart. It proposes already ideas to close the deal. And then when I visit the country, everybody is mobilized. Everybody has been working over the weekend. Vanessa is crazy running around, and we close deals faster.

Yes. Very exciting.

Thanks, Kirk. Sophie, we ill take our next question.

Our next question comes from Alex Zukin with Wolfe Research.

Congratulations on a fantastic quarter. Marc, you're seeing a pretty big debate intention in the industry about open weight and open source models versus frontier model adoption. I was wondering if you'd set the stage for us because what you're seeing in customers is important as you're offering the best of both worlds with both Claudeforce and AI Force/Headless. So maybe talk about how you see this debate selling and about the monetization opportunity for both Claudeforce and Headless and kind of when we should start thinking and seeing it in revenue growth?

Yes. I think a lot of customers are asked these questions, and there -- a lot of us -- I'm sure you can relate live in this podcast universe. We just were with David Friedman (sic) [ Friedberg ], whether I think we all love listening to his podcast, right? And we like listening to a lot of these podcasts. And we're in this podcast world, but I think sometime and please don't quote me on this, but podcast do not always meet reality. And it's very interesting. It can be very aggressive. It can be very stimulating, but it might not be exactly true. And I think that, that is just the world that we live in. We live in a world of misinformation where everybody has an agenda, financial agenda. They're trying to move a market. They're trying to make things happen. You see that as well with shorts in your market, where you'll see people come in with all kinds of misinformation to try to get something to happen. So I think that 1 way to think of -- to handle that is to go talk to customers. And this is what I do. So I have spent a lot of time with customers. And customers, it's pretty simple. They have -- if you talk to customers, what they'll say is, we only have 3 platforms. We have used Salesforce doing the front office. We have SAP, they're doing the back office. And we have Microsoft, they're doing productivity. And that's it. The reality is, yes, they have those 3 platforms. They also probably are hanging out with the hyperscaler. There's probably 3 or 4 different ones depending on what country it is in Europe. They also probably have a data lake supplier, and they also have a collaboration supplier, Slack or Teams. And I think when we look across those 6 things, they do not want to also then say, I'm have all these models, okay, which kind of gets to your question, which is like, do you think that these customers then are going to all of a sudden set up, all and maintain and build the human expertise to be able to build and maintain and train models? No. I believe what they're going to do is consume AI through packaged software. Not a huge statement, just they're consuming their AI through packaged software. You're using Slack, you're consuming AI through packaged software. You're using Salesforce. You're consuming AI through packaged software. At some level, when you use Cowork, you're consuming AI through packaged software. And when you're using Claudeforce, you're certainly using AI through packaged software. And I think that's the right way to think about it. I think the model, it's like, do you even know what chip you're using? Do you really know what data center you're on? Do you know what router you're on? What switch, what type of fiber what cable you're using, what interconnect you're using, at some level, we do will get abstracted back to kind of how the way the world used to be, which is we operate at a certain level of value and I believe the highest level of value is at the application layer.

So I couldn't tell you from my phone all the way through Salesforce and Slack all the way down to what the hardware infrastructure is, and I shouldn't have to worry about that. But I do understand this layer up here. And those top 4 layers that I've articulated several different times on the call, that's where I live with the customers. I did not have 1 customer really say to me, tell me, are you going to be an open source. Are you going to be on Frontier. And I think like 1 way to think about this is, yes, this is a highly dynamic market. And there will be frontier models. There's going to be open source models, and there's going to be room for all of the above. And I think it's very exciting if you're not tracking what NVIDIA is doing with Nemotron 1.3 and the ability to fine-tune it. It's a very exciting moment for models. But Salesforce also makes a lot of models, too, by the way. We have a lot of our own models that run in our platform. Our platform runs on our own custom models we've been writing models since like 2015, and you can find them all on hugging face. If you go, you'll see all the models. If you're really in the models, you can go and see that. So I just think we're at a moment where be careful and I'm not sure -- I don't think that's what customers -- well, I don't think that's where customers' heads are at, at least when they talk to me, that's not where their heads are at. Maybe some customers are. Obviously, if you go talk to my engineers, maybe they're interested in what's going on, but I don't think the customers that I speak to are at that level. Does that help you understand how I think about it?

Super clear.

Thanks, Alex. Sophie, we will take our next question.

Our next question comes from Elizabeth Porter with Morgan Stanley.

I wanted to just circle back on the strong Agentforce ARR growth. And in particular, what proportion of current Agentforce is coming from paid production customers rather than pilots? And how is the time line from pilot to pay deployment and refilling credits changed over the last couple of months? For instance, are we seeing any sort of lengthening of customers are evaluating a broader set of solutions or sales force is increasingly open and headless posture, helping derisk some of these decisions to potentially accelerate the agent with the Salesforce platform.

Can I take it?

Yes, I'm going to have to turn it over to Miguel. But the first thing I'll tell you is we offer our customers a wide variety of capabilities, products and functionality. And you mentioned Agentforce is 1, Coworker, if you haven't seen Coworker, it's our fastest-growing AI product. We have another super fast-growing product, Slack bot, where we expect Claudeforce to be growing super fast. Agentforce also is growing super fast. We have a lot of them. And what we're trying to do is let customers have very flexible capability so that they can choose all of them and grow with them as they want. And we have these new bundles that you've already heard about and these new bundles let customers do that. See customers want to buy and want to price in different ways. This is something I've learned really aggressively recently. Some are still buying by user and by agent. And that's important for them. Some of them want it might consumption, and that's important to them. Some of them are just basic usage customers, and they want to pay that way. Some of them even want to pay by outcome and that outcome could be by a transaction outcome or a business outcome. Because customers want that kind of diversity and pricing, we've created a high level of flexibility and that, I think, has really expanded our ability to sign very large transactions with our customers.

Miguel, do you want to go into the details?

Yes. I mean the top line numbers are obviously very impressive, but I'd like to go a little bit deeper under the hood. When you look to asset bookings, not just a cumulative ARR, our bookings, Elizabeth, grew double digit -- triple digits. So we doubled year-on-year. That was pretty fantastic. 50% of the bookings came from customers refilling the tank. So they consume, they use their flex credit, they want more, they raise their hand, we go there, which, by the way, is good for a productivity because these cell cycles are very short. The time to production is very important for us because consumption is very important to us. We bring customers in the funnel. We want to bring them down. The first thing is we get into production. We added 2,000 paying customers into production, that's 70% more quarter-on-quarter. We are not happy with production only. We want them to really be ready to do the work. We are taking them to the enterprise. We show them everything, how to work, and then we live. We call those agents consistently consuming. We added hundreds of consistently consuming agents down there. You've seen on the funnel. You've seen the consumption number that AWs exploded. It's -- but for me, the most important thing, the most important thing is to see the live customers. I want to tell you 1 story if I have time on 1 customer. But if you want to write down names of websites that you can go right now and see the agents or call them, Bodewell, geappliance.com, lululemon. If you're going to Mexico, volaris.com. Or Lingus, if you're going to Ireland, my 2 daughters are in Ireland now. You want to call Amazon Blink. I mean, you want to call [indiscernible].

Amazon Blink is such a great example, Miguel. Will you just talk -- tell the story just briefly?

I actually have the phone. Yes, Amazon Blink basically, if you have anything going on in your Amazon devices at home, you can call -- voice call. But I'm actually giving you 1 because since everybody hears a business, I want to -- if you need business insurance, commercial insurance, you can call 1 of the Berkshire-Hathaway companies, [indiscernible] and I'm actually going to give you the phone number, 402-200-3104, just called -- so 402-200-3104, and the experience is incredible because it's not just a chat, is a voice that is anchored to your trusted context and that can execute across our deterministic applications. But anyways, I give you a bunch of websites, et cetera. But let me tell you 1 story of some of the agents that are working in the enterprise that people don't see. Big digital platform company here in the U.S., 45 million, 45 million AW use is 1 of our largest, grew 14x. We started this 6 months ago in Q2. The AW user consumption grew 14x, 14-fold. And what it does is they call it the activation agent. They go to all the merchants that are already registered with this platform, but they are not transacting. They are not getting money from the merchants and the agent has 1,500 interactions every day, getting these dormant merchants alive and then kick in and then generating revenue. This customer, for instance, was 1 of the customers because we're meeting them where they are with pricing. They didn't want to buy upfront and do a big thing. They just wanted to try and see the value. So now, Marc, we are negotiating either -- we've given 2 options, either outcome by deal. I mean we're talking -- I mean, this is a $40 million customer for us. So it would be a master deal on [ adcom ] based or it could be simply an [indiscernible], unlimited agreement with millions of...

I think that's the most important thing for us. I mean, I think that we've heard so many companies say, we're doing outcome pricing. We're doing this pricing. We're doing that pricing. We're just too big. We have to do it all. And so we've built these new flexible pricing schemes that let our customers choose the price that they want. And we're going to get more and more into that zone. But this last thing that Miguel said that some customers want business outcome pricing, that's very exciting, and that's never been true before in enterprise software. And Miguel, 1 more huge announcement we had in the quarter was that Miguel is now our Chief Operating Officer, congratulations, Miguel.

Thank you.

Thanks, Elizabeth. Sophie. We will take our next question.

Our next question comes from Gabriela Borges with Goldman Sachs.

I think this follows nicely from Elizabeth's commentary on the different types of pricing models. If I think of what Zero and [indiscernible] is saying, you've got some customers that are using the Agentforce stack. We've got some customers that are doing Headless, and now you've got the Claudeforce announcement today as well. My question is a little bit on the economic value to Salesforce from these different types of cohorts. I know you're still figuring out monetization on Headless. Do you think in steady state, you'll be able to capture the same type of ACV from customers agnostic of whether they do Agentforce or whether they do some sort of Headless configuration? And then any early reflections you can share as you progress on figuring out monetization that this would be very helpful.

Maybe I'll start, Gabriela. It's a great question. And even going back to Alex's question as well. We feel really comfortable with -- like you said, we're working through Headless monetization. But just to further double down on Marc's point, what we ultimately think is predictability and flexibility is what's important to customers. And as Marc said, they're going to buy a lot of different type of ways I think most importantly, though, we think with our new platform, AI Force, we really are going to unlock the potential for knowledge workers in great ways. So I think that's going to be pretty impressive. The other thing is, is the way we're delivering this is through our premium upgrades that we also talked about. So they're going to get innovation in the moment. That is an early stage for us relative to our installed base. So our opportunity to continue to upgrade our customers to those premium additions that gives them instantaneous access to this innovation is an amazing future opportunity for us. So we feel really comfortable with the monetization strategy that we have and most importantly, the flexibility that we're giving our customers that allows us to capture that.

Quick thing. The AI opportunity, you can see, first, augmenting employees. That's where our premium additions come because when you are on 1 of those premium additions, you basically have a limited usage to augment yourself as an employee. And then the customer-facing use cases, which is Monster. This is the digital label world. It's Monster that we monetize with ILS with Flex credit and here, we meet customers where they are in their journey. Sometimes they want to buy as they go. Sometimes we just charge for overages. And many times, they buy upfront a band of credits. I think AI services is going to increase the demand across the board. What I can tell you is every and we're measuring this, we're going to give you an update in the Dreamforce, but we started last year. Customers do start the genic journey with Salesforce. On average, they are already at double the AOV with a perspective of nearly [ quadriplent ] 3x to 4x the AOV. Think about this. The only thing that I need to do is to convince every single customer, and that's why Mark is on the road from time to time, meeting customers and convince them to come like he did with [indiscernible] to come with us in their agenetic journey because once they make the decision, double, triple, quadruple the AOV with us.

Yes. And I think that, that's what's really key. I think your question is so good, Gabriela. I'll tell you why it's so good because I think that if you would ask us a year ago, we'd be like, well, we don't really know. But I think now what we see is, well, all this new action on the network, call it, agentic action, new user interfaces, new applications, is just a lot more activity on our service. I mean the amount of action at Salesforce on our service, the value that we're adding, the actual value that we can give to our customers is immense. So have we fully optimized that yet? No way. We're still trapped in some ways in old per user pricing models, but their opportunity to build much more aggressive pricing to really represent the value that we're offering to our customers, I think, is enormous. And Miguel is really at the very beginning of what I think will be 1 of the most exciting journeys he's ever been on.

Only 5% of the knowledge workers that use sales and service have upgraded to the higher-end additions. We get a 60% to 80% premium. So imagine.

Yes. And I think that's a major. So if you want Claudeforce, if you want Headless, if you want all these value-added agentic additions and capabilities, you need to move to our premium addition, which is what's going to give you the power to fully operate inside your enterprise and deliver this high level of capability and unleash the huge amount of value that's been trapped for so many years. I hope that makes sense.

Thanks, Gabriela. Sophie, we will take our last question.

Our last question comes from John DiFucci with Guggenheim Securities.

Miguel, I know you've gone through the math of how the subscription net new AOV to revenue works. And we agree, it's math. But there are 2 things that Robin said that I think are really important to that. And one, net new AOV growth is the strongest it's been in 4 years. And two, you had near record low attrition. Those 2 things help that math work. We know that. Can you better -- I guess, help us better understand what's happening underneath those statements? How stable or sticky is that foundation of your recurring revenue? And more importantly, how should we think about that going forward? Because Marc's talked a lot about the so-called SaaSpocalypse, and he's -- and I appreciate that those -- actually, I don't know, probably not fun, but it was always interesting. But if you follow through on those...

Thank you for your empathy.

But if you -- if Salesforce follows through on those statements and that math, both low attrition and strong new business growth, that term is going to be part of history, Marc.

You missed the third leg of the stool, though, John, you didn't say a new value-added for customers. I think that the key is 3 things. Yes, it has to be about low attrition, okay? And high net new AOV, very important. But the third thing that we have to do with this AI is deliver this new value added that's going to bam, bam, bam, offer all these new capabilities. And I'll just come back to what we said, whether you're using Claudeforce or you're going to see the new version of Slack. I'll just go into in a second, also has an incredible surface. And this idea that all of a sudden, these kind of things that sit kind of theoretically at the top of the stack, can dynamically build for you these applications. So you're going to work with it and you're going to say, "Hey, I'm managing this part of my business. It involves these customers, these products, these competitors, this technology and bring this together to me in a map and this and that." And where you used to have to take months or years to build very complex, very dynamic applications that look like they were built in. I'm not going to get too geeky on you but like react like a very complex user interface, all of a sudden, that is what the AI is really good at. And we're doing the heavy lifting down below the sharing models, the business models, the security models, keeping it safe, keeping it controlled, keeping the data where it needs to be, making sure the wrong person doesn't see the wrong data, that kind of thing. But then that idea that we can offer a lot more value. So it's those 3 things all coming at once. And that is what -- that's what our hope is, is that we will offer more value for our customers and unleased trap value and that, that's going to give us this huge leg up. And that last thing that Miguel said that we're moving to not just kind of outcome-based pricing, which is we completed this many phone calls, therefore, give us do. We want to be able to say, no, we improve revenue by this much, so give us $2 because we made you $20 or we made you $40. And I think that, that's like the next generation of enterprise software that is more than outcome-based pricing. And there's other enterprise software companies that are out there that are kind of doing well. They work with sometimes mid-market companies and get huge prices for their products because they've basically been able to write an outcome deal like if I can generate this much more revenue for you, you have to give me a percentage of it. Well, you could do that at scale now that you have this kind of thing. And you don't need a bunch of FTEs and engineers and all that to do that because that's what the AI is really good for up here at this top level. Does this make sense, what I'm saying to you?

You know what, Marc, it makes absolute sense, but -- and I really appreciate it. But I think you'd agree with me that even without that, the way Salesforce stock and some of your peers have been trading, they are not even looking at that. They were just afraid things were going to deteriorate into...

They're not looking at -- right, well, yes, take the value of Slack. Obviously, I get a call every day from somebody who wants to buy Slack. Hey, take the value of Slack. Then take the value of our Anthropic stock. That's been like half our value of our -- forget our cash flow or our customer base, that's why I said, Robin go try to buy as much back as much Salesforce stock as you can. And I think we got a lot of stock at a good price.

Well, you're on your way this [indiscernible] statement.

Thank you. Thank you. Keep guiding us, John.

On the confidence about the future, I mean, H1 -- H2 last year was our great net AOV growth semester. We knew, Robin and I, we partnered together, we knew that for organic revenue to reaccelerate as we committed to happening now in H2, we had to continue with UV growth outpacing AOV growth. That was our obsession. We aligned the whole company this fiscal year, compensation plans, every role. H1 net AOV significantly growth outpaced AOV growth. We've already raised the guidance. We've already committed to the revenue acceleration in organically in Q3. The momentum is there. We feel very, very strong. The demand environment is very, very good, very strong. Our pipeline at record levels. We have more opportunities to monetize that we are not even put it in the pipeline. We had the capacity. Thank you, Marc. You pushed us to hire a lot of AEs when things were a little bit rocky.

Somebody's got to call these customers besides me, Miguel.

I know. I know. I do some calls also.

Thank you for helping out.

We have the capacity and we have the innovation, we have the product. So do you think the demand is there and it's going to get even better with the AI Force. And the supply, which is, for us, supply is capacity, distribution capacity, which we have higher more than anybody else. We still have to hire 1,200 more before the end of the year. We are now hitting 15,000 AEs and the innovation. We have the best products we've ever had. Most of it organic. We've made some tuck-in acquisitions to enlarge our times. We are very confident. Definitely, we are committed to the H2 revenue reacceleration. That's already math. Q3 is math, Q4 is nearly math. But we are committed in sustaining that acceleration, and we are going to hear more at Dreamforce about fiscal year '28.

And I think even to add to that, Miguel, is if you take a look, particularly at sales and service, we're seeing durable growth in seats, right? So that's fundamental proof point is that everything we're doing is amplifying our entire platform. It's an amazing progress.

AI, It's simplifying the need for drastic context and deterministic execution. Very clear.

Perfect note to end on. Thank you, John, and thank you to everyone for joining the call. We will be hosting a webinar on Tuesday to deep dive on our customer motion and adoption of AI with Bill Patterson and Conor Marsden, and we look forward to seeing you all over the coming weeks and at Investor Day at Dreamforce.

Thank you for joining. This concludes today's call. You may now disconnect.

All right. We're going to get started with our next session. Very pleased to have Patrick Stokes with us for our next fireside chat. Patrick, of course, is President and CMO of Salesforce, has nearly 20 years of experience in product leadership roles. That includes the last 15 or so at Salesforce. Patrick, thanks so much for being here.

Thanks for having me. Really appreciate it.

My pleasure. So just by way of getting everything started, it'd be helpful if you could outline for us your primary responsibilities and focus areas at Salesforce.

Sure. Happy to. Well, I'm exactly who you all want to talk to. I'm the Chief Marketing Officer at Salesforce. I've had a bit of an interesting journey. So I've been at the company for about 15 years, longer than most, not quite as long as some. Most of my career has been on the engineering and product side. I started as an engineer, moved into product kind of 10 years in product at Salesforce. I ran our platform, and then moved over to marketing about 2 years ago, 2.5 years ago, right, when all the AI stuff really kicked off, I didn't quite know it at the time and then the whole world changed, which was difficult. And then about 4 months ago moved into the CMO position.

Fantastic. So you're coming off of a solid Q1 in which Salesforce impressed with Agentforce disclosures, of which there were several, and you reiterated confidence in second half revenue acceleration as well. So maybe just briefly recap for us what you personally viewed as most important or instrumental kind of out of the last quarter.

Yes. Well, certainly, we're very excited about the continued growth in Agentforce. I think $1.2 billion now on the Agentforce side, up from, I believe, $800 million. So we're very, very excited about that. We're starting to see customers reach real scale on that with some pretty sophisticated use cases. I'm also equally excited about the AWU growth. We're certainly seeing our own token consumption from.

And that's agentic work units.

Agentic work units, yes. And this is this idea of tokens in and out, that's effectively a measure of reasoning or a measure of intelligence, but it's not a measure of actual work getting done. And so we thought it was important for the market and really -- and the technology sector as a whole to kind of have a way to measure actual work getting done by these pieces of intelligence. So we introduced that in Q4 and I was very excited to see that continuing to grow alongside token growth, of course. Interesting watching the 2 of them kind of grow at -- each of them kind of picks up pace at certain moments in time.

It's also really interesting looking at the AWUs kind of across different industries and different segments, you can start to see where different usage patterns are emerging. And then lastly, I would say, I was very, very excited about our Headless launch, which we did not that long ago, in March, I think it was at TDX, our developer conference. And I'm very excited about what that means for Salesforce and the reaction to it we're seeing from our customers. So really opened up kind of a new way of thinking about our role in this agentic era.

All right. That's great. I definitely want to dive more into Headless in a few moments. But let's stick with Agentforce because I think it's important to understand what's underpinning the growth that you were just talking about. So there were a lot of announcements around Agentforce, I would say, dating back to Dreamforce last October. Many other improvements have since been unveiled as well. But how would you characterize the agentic capabilities of Agentforce today? And what are the most important enhancements to both the tech and the ecosystem, right, over the last 6 to 9 months that are really enabling your customers to unlock more value?

Yes. I mean I think the 2 biggest advancements in the last, let's call it, 6 months or so on the Agentforce side are one, something that we call Agentforce Script, which is, if you start building an agent, you get very excited early on because you're, like, all I have to do is write human instructions and like I've coded my agent. They're effectively -- if I'm being very reductive, they're effectively just prompts under the covers. And so anybody can write those. But what you realize when you're starting to try to get agents to do multi-step and complex workflows is you have a lot of like do this, but unless this happens, then do this and you start writing it like that and the agent gets very, very, very confused very, very quickly.

And the irony of like programming languages is that's effectively what they are, if this, then that types of flows. And so we pioneered this way to kind of put little micro moments of scripting into the prompt so that you could eliminate some of the probabilistic problems of working with an LLM. You've all seen this, you ask it a question and then you ask it the same question, you get 2 different answers. So it's a probabilistic nondeterministic system. When you're trying to execute workflows, you don't want that, you want deterministic. You want do it the way I've asked you to do it every single time. And so Agentforce helps you with that. And that's been a big unlock for our customers that are trying to either do this at scale or do it in regulated environments, do it with certain policies that they need to make sure the agent is following.

And then the second, I would say, is voice for sure. Voice is very, very exciting for our customers and it was a pretty significant computer science project for us to get that working well. Voice is a tricky thing because there's humans and people have different ways of talking and there's interruptions and latency and all of these things that you kind of don't think of until you actually start trying to build it and you're like, oh, this is actually pretty damn hard.

Yes. And there are, Patrick, several CCaaS incumbents, right, that have had existing solutions for a while. How do you think about where Agentforce voice stacks up at this stage?

Yes. I mean, I think we're in pretty good shape. If you had asked me 3 months ago, I would say we probably have work to do. But I think at the moment, we're now running Salesforce, our 1-800 number on Agentforce voice. You can call it now and you'll talk to Agentforce. And we have other customers doing the same. I think over -- we'll probably be sitting here with our Q2 earnings after talking about the big kind of scaled voice customers, just like we are the kind of chat customers with Agentforce.

Okay. Super interesting. And then one other comment on Agentforce. We started to hear of some forward-thinking customers that, I would say, are deploying multi-cloud Agentforce use cases. So they might extend from Service Cloud to Sales Cloud or you're doing case resolution in Service Cloud and that triggers a campaign in Marketing Cloud. Is this something that you're really kind of seeing as well? And is your go-to-market aligned enough, frankly, to be able to sell more holistically in this way?

Yes, good question. I mean it's funny. The first part of your question is like, we saw that from the very beginning. There really isn't a usage of Salesforce that doesn't cut across. The platform is actually quite a bit more like [indiscernible] together than you think that it is from a pure usage perspective. From a buying perspective, it's maybe a little bit more discrete. You have sales service, et cetera, et cetera. So those -- as soon as you start using Agentforce, you immediately get into scenarios where you're like, okay, I'm going to use it to try to qualify a lead. But then I want to put those leads into a marketing campaign. You just -- you immediate -- or service, I'm going to use it for case resolution, but I need to know if there's an open opportunity because I'm going to handle my case resolution in a different way. So there's tons of those types of use cases.

What we've tried to do is make it easier to buy that. So this is what our kind of top end additions like we call it internally A for X, so Agentforce for sales and commerce service, et cetera, they kind of come with the entirety of the platform. It's like here it all is, go implement your agent and so you're not kind of buying individual piece parts.

Okay. Great. And then maybe just to sort of zoom out a bit, I think what a lot of investors in this room and elsewhere are wondering is like when will the rubber hit the road, right, when it comes to enterprises, meaningfully deploying Agentforce in production and at scale. Certainly, we know that some have gotten to that level, right? But it just comes down to more broad adoption in that capacity.

Yes. I think your second part there kind of got it. I think certainly, we are seeing many, many, many hundreds of customers reach meaningful scale, either scale from a usage perspective, in some cases, simple use cases, but that are being interacted with by hundreds of thousands or millions of individual kind of consumers on the other side, customers like Southwest, who are now kind of doing 20% to 30% of all of their inbound requests from customers with Agentforce. So it's like the curve of sophistication of the interaction and then volume. And so there may be kind of lower on sophistication, but very, very high on volume. And then you have other customers that are very, very high on sophistication and maybe a little bit lower on volume as they experiment.

So we're seeing kind of all ends of that spectrum. But I think the real basis of your question is, okay, Patrick, but like when is everybody going to be doing that? When is [ 187,000 ] of your customers, they're all going to be doing that?

Or at least a very significant percentage.

At least a very significant percentage. I think we're definitely nearing that. It's certainly part of how we think about H2 and accelerating the business. That's kind of all factored into that. So I definitely think you're going to start to see a rapid acceleration as the product gets easier to use, as the CIOs start to kind of trust what we've built and that also means in a way, I have to be careful how I say this, but in a way, given up on what they've tried to DIY, right? It's kind of like the cloud in the early days where everybody went out and tried to build their own cloud. And you could look at it and say, what are you doing? And what they're doing is they're creating intuition of what it takes to do it.

And then that's putting us in a better position because now they come with that intuition and they go, okay, you've actually solved the problems that I ran into and maybe didn't want to solve. And so we're definitely starting to see a lot more kind of openness to the platform from CIOs than maybe we were 6, 12 months ago when everything was so new, they were all just trying to figure it out.

Okay. Yes, super encouraging. And of course, just maybe to sort of stay with this train of thought. I mean the big fear out there remains that AI and specifically the Frontier labs, right, that they will drive significant disruption and deceleration in the Salesforce business. Why is that view incorrect?

Well, I mean, there's the data side of it, and then there's the subjective side. I mean the data side shows that it's not correct at the moment. Our seats are still growing. Our businesses -- our core businesses are still growing, and we're very happy with the growth. But also when you look into the AI labs, this is what's most fascinating to me. And many of these labs are run by -- have people in the go-to-market organization that we know, they're all using Salesforce extensively, in fact, more than some of our biggest customers. And the reason that is, is because these AI labs, what they're doing is they're not using Salesforce the same way customers have for the last 20 years.

They're not using it as a UI. They're not logging into it every day and logging a call and an opportunity. They're using it through their own agentic interface through Claude Enterprise or through Codex. They're hooking it up with MCP. This is how Headless comes into play. And that's how they're using Salesforce. So what we're seeing is there's actually an expansion of usage and expansion of consumption. That's separate from the whole seat conversation is kind of a separate one, which I'm sure you'll hit on. But just if you separate that for a moment, the existing seats this new way of working with Salesforce, we're seeing usage spike up quite a bit. And so that's very, very encouraging for us.

Okay. Terrific. And yes, that was a perfect segue to Headless. So I do think the Headless 360 announcement, Patrick, just seems like a really clever and interesting way to drive stronger connections to your point, with the Frontier AI models, also opening up the Salesforce platform to external AI agents and coding tools via MCP as you also highlighted. One other thing I would add is just also reducing friction, just making Salesforce easier for developers to use. So I guess, is there anything -- is that a fair characterization? Anything else that you would sort of highlight as part of this?

Yes, definitely a fair characterization. I mean the inspiration for Headless came from just watching people use Salesforce in a different way. We have these partnerships with the AI labs and we're watching them use and we can see the API data. We can see that their consumption is through the roof, but like they're using it through Slack, for example, the AI labs are all big giant Slack users or they're using it through their own Claude interface. But we are also seeing it just in the -- this is a very, very hot kind of space right now. So you go to X or you go to Reddit, and it's just like an army or you go to these Discord channels where all these developers hang out. And there's just these constant kind of sets of conversations of people trying to figure this out.

There's something that I think people can sense where they're like, there's a new way to work. I think that's what they're sensing that there's going to be a way to work that all of the friction of going to all of these discrete applications that have been purpose-built for the function that I do, that's going to go away and it's going to be replaced by a new interface that interacts with kind of the underlying capability of those discrete applications.

So the applications aren't going to go away, but the UI is going to be massively disrupted. And so we started to see that pattern emerging, and that was the inspiration. It's just like, okay, well, that's what customers want and many other companies kind of in our space saw it as well and they were like, we don't want that. We're terrified. We don't want to lose our users. We don't want to lose the UI, and we just took the exact opposite approach and said, no, we endorse it. Let's open this up. We're going to have to figure out how to monetize it, which is something that we're talking a lot about, and I'm sure is very relevant for all of you, but the pattern itself is really a no-brainer. And that's why I think it's so exciting.

I think that you saw a very positive sentiment, which we were all very excited about. It's been a couple of months of the SaaS [indiscernible]. And so kind of getting back in front with a message of like, no, no, guys, we see this. We understand what it is you're trying to do and we want to actually enable that and endorse that. We think that, that is the future. I think it was a little bit of a reset moment for us.

Okay. And clearly, it sounds like a mechanism that put Salesforce more directly in the token path.

Absolutely. Well, I mean when you look at these tokens, tokens represent intelligence, as I said earlier, and right now, by far, the single biggest use case for consuming tokens is coding, right? That's -- it's the killer app. One engineer can generate like $100,000 a month bill without much problem. Now you can debate whether that's highly efficient or not. But they're making -- these labs are making a lot of money on these tokens.

There will probably be some sort of normalization or reckoning of that. Right now, everybody is just like everybody code. And so people are -- you hear these stories of people taking their -- blah, blah, blah. So there's going to be a normalization. But what the AI labs are -- we think that they're looking for is, well, what's the next killer app? What's the next killer use case? And that's where we think Salesforce is perfectly positioned because we think it's knowledge work. We think it's people like you every day that are showing up and having meetings and analyzing things and making decisions, you have to access information and you want to be able to access information in a way that's low friction, and then you want to be able to connect that to the intelligence of the AI, and that's effectively what we can provide.

Okay. Terrific. And then you alluded to this, Patrick, we know there has not been an official decision on Headless 360 pricing, but maybe you could just sort of speak to some of the early considerations and how that monetization might present itself for Salesforce.

Yes. I mean, so there's kind of 2 arms to the puzzle. One, there's actually quite a few very important technology decisions that we need to make in terms of how do we ask people that are building agents outside of our platform to identify those agents to us in the same way that they identify a user to us, and that's important to solve because that provides the layer of kind of governance and licensing and permissioning that we need to put for the agent to exist and consume from Salesforce in the first place.

So there's a number of technology decisions there, many of which will almost certainly result in some sort of agent user license showing up. So just like we have human licenses, we'll likely have agent licenses as well where you have to self identify the agents that you run on top of Salesforce.

Now that sounds like if you're in Salesforce at least where we have this 25-year legacy of seat-based pricing that sounds like the answer. Okay, great. So we're going to charge for the agent licenses, that's possible, but we're trying to be as thoughtful as we can on this and make sure that our own kind of legacy bias on that doesn't come in too much. We're trying to be very kind of forward in the way that we think about it. But really, what we're doing is we're talking to our customers and our partners. And we're going to them right now and we're saying, look, blank slate, here's your contract. Imagine that you could just rewrite your contract right now and have whatever unlimited usage of Headless that you want for the environment that you're trying to create. You tell us what that contract would look like. And that's -- those are the conversations we're having with our customers right now to make sure that we can do it thoughtfully.

What we don't want to do what you can feel in the room, I feel it in every room and especially rooms like this is you all want to know the pricing model so you can model our future growth. And we want you to be able to do that as well, but we want to make sure that we don't give you something that turns out to be wrong and then we have the wrong model. So that's why we're being careful here.

Yes, it makes perfect sense. And the other thing, which I don't know how much of this you guys have thought through, right? But if I think back to the early instantiations of Agentforce pricing, right, where -- and this is a very different world, right, because everyone was trying to figure it out.

Like 6 months ago.

Everything is [indiscernible]. So -- but initially, it's per conversation pricing and then it was per action. And then we got to the point of having sort of discrete subscriptions, right, then we have the Agentforce 1 Editions and then you had AELA, right, where you have Agentforce ELAs for customers that are willing to make very, very big commits. So I guess the question here is, is the bias from your perspective, and I won't hold you to it, I realize it's early stages, but to sort of give customers some choice, but maybe to not make it overly complicated, which some may argue was maybe an initial impediment to Agentforce adoption, again, in the very early innings before you kind of were able to work through all that.

I think that's a very fair characterization or criticism, whatever you want to call it. Yes, I think there's like a Goldilocks type of scenario that we haven't totally found yet. When we first started with Agentforce, it was something that was brand new, and customers didn't know what to expect. So they wanted a consumption model. And so we gave them a consumption model, but it turns out that is a very complicated consumption model. And so it's very, very difficult to kind of predict. It was hard enough to predict their own usage because they didn't really know how they were going to use it at the time.

And then even if they could predict their own usage, it was hard to turn that usage into to understand the commercials of it because our consumption side was so difficult. So we made it easier. We're like, okay, what if you just bought the AELA and then you don't worry about it, like use as much as you want and don't worry about it. And customers like that as well because it's simple, but it's also very expensive. It's more expensive than -- so it's -- you kind of have to pick your [ evil ] but that's not what we want. We don't want the customer to have to pick their [ evil ]. We want to get them to something that they can really trust and believe in. And I think we're seeing other models emerge outside of Salesforce that are interesting. And I think what this comes down to is it's not really about what benefits us. It's we want to find a model that benefits our customer. And so we're going to experiment with as many models as we can. The downside to that is it looks like we're confusing the market, which I get, like how many pricing models do you have? And how do we measure this? But we're -- our approach is like, yes, that's a moment in time, and we kind of just need you to trust us on this. We're going to figure it out. We're going to do it with our customers, and we're going to find the right thing to do.

Okay. Very helpful. Yes. Thanks, Patrick. And while we're on the topic, so last week, Salesforce announced the acquisition of Contentful, which has a CMS content management system. Can you expand on kind of what this IP will help Salesforce accomplish?

Yes. Well, first of all, so the interesting thing about Contentful is that it's effectively a headless-first platform. So they didn't spend a great deal of time worrying about what the UI for content could look like. Because if you think about content, it's really just like a feature of a campaign, which is a feature of marketing. So the SaaS era has created just like this massive sprawl of these purpose-built applications, and that's especially true in marketing. If you look at the LUMAscape for marketing, it's like there's so much. So they were like -- we don't think that's where the world is going. We think the world is going into a world where there's some sort of intelligence that's orchestrating campaigns in real time and doing one-to-one personalization. And so what it would need to do that is it would need a headless CMS. It would need to be able to pull the content out when it needs it. The CMS would have enough metadata, enough context in it so that, that intelligence on the other hand can go grab that when it needs it.

So part of the attractiveness of the acquisition was certainly that just bringing a little bit more Headless DNA into our product organization. But also, this has been a little bit of a gap in our Marketing Cloud Strategy for many years, and so it certainly had some attractiveness in terms of kind of shoring up our Marketing Cloud.

All right, terrific. I'll ask one more question, and then we'll pause for any of you folks that may have a question to ask Patrick. So I want to just talk about Slack for a moment. So a lot of investors, quite frankly, believe that Salesforce overpaid for Slack all those years ago. Now my opinion is that view has some merit, but I will also say that Slack has quietly become a more integral technology in an agentic world and many people don't really seem, from my investor conversations, really seem to be aware of that to the extent that it's actually occurring. And going forward, it also seems that because of Slackbot that this can really become more of a center of gravity for Salesforce. That's the opportunity that we see, but it will be helpful to hear your vision on where you think all this is heading.

Well, it's pretty similar to your characterization there. Slack for us, I think the market in general kind of see Slack as a collaboration tool and they kind of put it in the same [ bar ] as Teams. And I think that's just a really kind of unsophisticated view of what Slack really is, which is where Slack started was it was a platform for developers. So developers needed a way to work together to build projects and so Slack emerged as a solution for that. And it wasn't just because you could create channels and invite your friends into the channels, it was because of its openness. It's connectivity to things like GitHub and Jira.

It built this ecosystem of connectivity. So it became very, very sticky for developers in the way that they work.

We're seeing that same thing now happen to all sorts of different functions. Certainly, this is true in Salesforce. I mean we only work in Slack. It's unbelievable how much we get done in Slack. But we're seeing it with our customers as well and especially the small -- the new logos, the small customers and even the AI labs, which we say small, they're like $1 trillion companies, but they only have like 2,000 people in them, right? So on the scale of employees, they look like they're small, but like they run their whole business on Slack, and it's because of that connectivity. So that's kind of one side of it.

But the other side of it, that's very exciting is that it is already a conversational interface. So it's already really well prepared for an agentic future for when you have teams of humans working side by side with agents. But it's even more than that because you could say, well, Patrick, we kind of already have that with Claude and with ChatGPT, that's a human working with an agent. And it is, but it's one human working with one agent. It's a single player environment. And that single player environment, it's useful. It's nice to be able to work like off on the side and just talk to Claude, but most of us work within teams.

Slack is inherently a multiplayer environment. And once you start putting agents into a multiplayer environment, all sorts of interesting kind of usage patterns start to emerge. So one person can ask a question and the agent answers and then a second person can ask a question and it becomes this very interactive multiplayer type of experience. Imagine using Claude Code, which is also a single player environment in a multiplayer experience. These are things that the labs are kind of very interested in doing with us at the moment. And it also adds an element of trust.

How many of you have ever had your boss and your managers send you something that you can tell was just done in Claude and they're like, look, here's my analysis. And like you instantly know it was done in Claude just by the way it's written. We all have a spidey-sense for this now or OpenAI as well. But what you also are probably developing a spidey-sense for is that your manager is [ full of shit ] and whatever they prompted to get that answer was full of bias. And so of course, it's very convenient that the analysis that the AI did matches what the person on the other side is asking for because that's just human nature in the way we prompt it.

Well, when you move that into a multiplayer environment, all of a sudden this kind of new paradigm of trust evolves because now you can see the prompt, right? And you understand how it arrived at the answer. And in fact, multiple people can kind of prompt it at the same time. And so you're getting a much richer answer that's much more based in multiple viewpoints, which is a much more trusted answer. And that's something that I don't think anybody has really demonstrated yet, but we're on the path to do.

I have to say that's a fascinating point. I hadn't thought through that in that degree of depth, that may be something again that really enables you to kind of drive some more separation, right? As Slackbot continues to -- and Slack more broadly continues to get adopted. So that's -- yes...

And even Slackbot, amazing tool, incredible capability, but even Slackbot right now at this moment in time is a single player, right? So you ask it a question, it answers for you. You can share the answer. But imagine, instead of that, you're just in a channel and you have 15 members in the channel, but 2 of them are agents. And you just ask a question in the channel and the agent responds to them. So that user interaction model or paradigm hasn't been fully explored yet, but that's where we're going with Slack and it's super exciting.

Okay. Tremendous. So fascinating conversation. So that took quite a bit of time, but it was important, I think. Any questions for Patrick.

Patrick, thanks for sharing the insights. Multiplayer was [indiscernible] as you think for multiplayer environment, essentially sort of like [indiscernible] system [indiscernible] use all that stuff. How do you think about that?

Well, I think what we have at Salesforce, whenever we're building product and trying to bring product out into the world, what we're trying to do is identify what's the thing that's scarce. Right now, the intelligence is not really scarce. I mean there's limits on how much energy we can provide in the world to kind of deliver the inference through the GPUs. But the thing that Salesforce delivers that scarce is context and trust. And so if you're building anything with AI right now, it's very, very easy to just start talking to an LLM and you get excited about how intelligent the answer sound, but then you realize it doesn't know anything about your business. And so then you start trying to figure out ways to dump information into it, so it understands your business, but that runs into all sorts of complexity and limitations in how much you're putting into the context window.

You can get about 1 million tokens into a context window right now. But if you put 1 million tokens in your context window, that's going to be an insanely expensive call that one question is going to generate a ton of tokens and the more you put in there, the more confused your LLM will get. So you have to engineer this way to get the exact context that it needs in the moment, pick the right model and then deliver an answer. And that's what Salesforce can kind of do behind the scenes in that moment. And what gets really interesting in that multiplayer environment and inside of Slack is the one piece of context that nobody is really truly thinking about, well, there are some us and others that we're working with. But the really interesting piece of context is the institutional knowledge of the people in your organization, which is encoded as those conversations inside of Slack.

So we bring that in as well. It's not just your data and your metadata, it's the conversational data as well, which adds all that additional context, and then we engineer a way to bring that all in at the moment that a question is asked, optimize the token spend, and that's really where we think our value is going to be in the future. Hopefully, that answers your question.

30 seconds if someone has a quick one for Patrick. If not, I'll just ask one more. So I just wanted to bring up a recent podcast with Marc Benioff. He mentioned that Salesforce is on track to spend $300 million on tokens from Anthropic this year, a pretty massive amount. Tell us just briefly what that will do for Salesforce in terms of productivity and innovation and maybe even inclusive of the cost implications.

Do you want to see our leaderboards? I can pull them up on Slack right now. We can go through team by team and see who's really using. Yes. I mean it's just been -- honestly, it's been insane in our engineering organization, watching the pace of innovation right now is unbelievable. And there are pockets where it's even more unbelievable. Slack, for example, is just fascinating to watch that team now.

I show up with ideas. And like on Monday, the ideas are in production, which is really -- it's just astounding to watch. But it's going much deeper than that as well because a lot of that -- most of that spend right now is coding, but not all of it. There is a very -- this is what I don't think the market has totally caught up to yet. A very material amount of that spend is just knowledge workers. It's my team in marketing. It's Val's team. It's people in sales. It's people like Miguel that are using it to do their forecasting. They're not coding. They are hooking up the MCP servers and consuming tokens from Anthropic to do their knowledge work. And we think that, that's where the kind of next big wave is going to come from.

Super insightful conversation. Patrick, thanks so much.

Thank you.

Thanks, everybody. I assume we'll have some folks filtering in after lunch finishes up. But I'm super excited to have Parker Harris with us, Co-Founder and Chief Technology Officer of Salesforce. Just a couple of English majors talking about headless technology, should be good.

So really excited to have you here. So much going on in the industry around agentic, AI, you've been through so many of these cycles. So it will be fun conversations.

Never been through a cycle like this one, but I've seen the cycles.

No. No, I don't think anyone has in terms of the pace and [indiscernible] and just sort of size of it. It's pretty amazing. So why don't we just jump into it. And if you have a question, raise your hand, we try to keep this as interactive as possible.

But -- to your point on cycles, you've been through a bunch. Why -- and we've seen a lot from Salesforce over the last month or so on Headless. Why are you all as a company sort of excited about that as part of the broader AI strategy?

Yes. I think we were surprised that we didn't make it the headline of Dreamforce last year. It was kind of a more recent idea, and we launched it at one of our world tours. And the feedback was just phenomenal, like everyone in the press and then on social and our customers are like, well, this is a brilliant strategy.

And I think what we're most excited about is just meeting customers where they are. We've had APIs to our service forever. But with the rise of -- and it's also kind of related to Claude Code that really hit that tipping point in February, that the first place we thought is Salesforce should just be easier to configure, to implement, to diagnose and why not vibe code it. So that's like the first step. Like let's open everything up Headless, and you can hit it with that. But then if you look at. And Salesforce always follows consumer trends. Like when we started the company, it was about Amazon, the bookseller, when we launched Chatter, was looking at Facebook. And right now, you look at the model companies and commerce, and there's UI coming into these products.

And so part of Headless was also, let's rethink our experience layer, the experience is actually in the Headless layer because you define the user experience and metadata. And we interpret it and we play it out in what we call Lightning. Now that can come to you, but you're not saying what I told you, you should see, you're just telling AI, this is what I want. And so give me my top deals for the quarter. Tell me what trouble tickets or cases that Kirk might ask me about at Evercore, and it paid in that response, beautiful UI, not just a bunch of text. And so it's really the new experience layer. We're seeing customers use it from things like Claude Cowork with OpenAI, ChatGPT, but also from Slack, which I've been spending a lot of time with the past couple of years being a great engagement layer for kind of everything Headless, not just Salesforce but everything in the enterprise.

Okay. We'll definitely talk more about Slack. I guess when you think about the Headless strategy, what does success look like? Is it opening up the TAM again for you in terms of just these people that might not have come through Salesforce traditionally through, say, more of the app layer? Is it -- when you think about where you'd want to be in a year on this strategy, what would you guys think about as success?

I think first and foremost, it's about adoption. So users are moving and they're looking at these new services -- surfaces. Success would be massive adoption of Headless. And I haven't seen the stats of MCPs on the Agentforce side been more close to Slack. But the Slack business unit MCP interface just spiked, we just released it, I don't know, a couple of months ago, and it is just spiked. So the number of people wanting access to that corpus of information is just like. So we're seeing the adoption.

And we're talking a lot internally about what are the monetization strategies for this because I think part of the success is also there is our current monetization like let's just get more license revenue and agent may be talking to Salesforce agentically through Headless, but it's talking as a named user because it needs to get the right data with the right security protocols, the right context for that agentic response. So it's a lot of named users. But there's also opportunities for usage-based pricing. And we're talking to our customers and saying, well, where do they want us to go.

Yes. That makes tons of sense. You mentioned Slack. Let's double-click on that a little bit. Probably one of the products when you think about Salesforce that has perhaps the most network effect to it within your customer organizations. How does that feed into sort of the broader headless strategy? Why is that such a great engagement layer for the agentic enterprise? Can you talk to us about that a little bit?

Well, take Slack where it was most successful when it started before we ever acquired it, was it with engineering. The engineers would take it real great. I'm going to hook it to get for source code control and Jira for my bug tracking and planning and connect it to my monitoring. Give me all the tools, but don't make me leave what they call the flow of work and just work there in context. And what is amazing about Slack is that expanded from engineering groups to all knowledge workers where they're working together and humans all humans at the time before AI, it's like, great, I can work in Slack. And we're just getting more work done. It wasn't just about communications. It was really about work.

Now it's about AI. It's about getting work done with AI, both as my assistant there. So Slackbot being a native one, but also third parties Claude Cowork in there or my Linear agent, if I'm coding, they're all in Slack because -- and they all want to be in Slack. And it's basically where AI-assisted or more and more AI autonomous work is getting done, but it's where humans are working together with AI with each other. And so Slack calls it multiplayer. When I use like a Codex or a Claude Code, that single player. I'm just working myself with it. And then -- but if I wanted to work with other people, Slack is really the best place for that. And so you'll see more things coming where we're opening up more surfaces where when people want to work together, whether they're coding or they're doing knowledge work, they're in Slack.

And by the way, in both Anthropic and OpenAI, like that's all they use, Slack. They have Slack, they have Salesforce. They don't really log into Salesforce because they're sitting in Slack using their models and stuff they've built, sometimes our stuff and working with all of these headless APIs to get their work done.

Has AI given you an opportunity to go back into those customers that might have bought Sales Cloud 10 years ago and say, look, like financial services is a good industry as an example. I guess, it's never been a great Slack industry forever reason. Maybe people are in Bloomberg.

Would you like to buy some Slack?

Our CIO is here. You can pitch them. But I think the idea would be you should rethink this in concert with AI? Is that kind of the message your salespeople are trying to reintroduce it to sort of, again, expand the surface area where you've been? And I guess, have you seen early success on that and maybe financial tough one, but other industries.

Well, let's take sales, for example. So we always try to like use everything ourselves first of all, call it [ dog footing. ] And so the sales manager agent, as an example, is this agent that is built on Agentforce that we have all these leads. We have a lead database, all these prospect leads. And there's a lot that we think aren't valuable, like to call them because you're too expensive as an employee to call them, call these because we think [ these are close ].

But we've taken all the leads. We think or lower value, and we've put them on this -- the sales manager agent, which agentically is having conversations with our customers over Vimo, WhatsApp voice is coming where they maybe call you. But it's not a one-way batch and blast like market automation like, hey, you're interested in Salesforce automation. And see if they clicked and then somebody calls it, it's a multiparty conversation back and forth. So that's an example where we can go back into a customer and say, would you like to close more business without adding more humans we can help you do that. Or the qualified great acquisition, ex Salesforce team came back in recently come to the website and just engaging with the customer on the website as an agent to get that prospect to the right place where maybe they'll even buy or they'll hand off to a human.

And so there's huge opportunities that we have just to go back on the customer base and say, are you in agentic enterprise? Have you found more productivity with AI or not? And if you've not done any of that, we can help you get there.

And you mentioned sort of monetization around the Headless list concept. I mean you guys had the app exchange for a long time, right, for API-based sort of revenue stream. I mean, should we sort of think about that in a similar vein, whereas you can still buy agents directly from Salesforce, you can build within Lightning? Or look, you might be able to build agents on Claude, hit, hit, come in through the MCP server and get data that way. Is that sort of the way we should think about it? And I guess, from your perspective, again, you're just trying to meet the customers where they are. Is that kind of idea?

Yes, so we have an agent exchange and agents can be built an Agentforce, a third parties. They can -- if it's built on an Agentforce, it's not going through MCP. It's just native or third parties can go through the MCP interfaces. Customers are building some themselves, which is totally cool. We're just trying to solve what is that use case they're trying to solve. And is it more sales? Is it happier customers in the service department? Is it lead gen and market automation whatever it is.

And we're going to do our best to provide services that I guess just you can do it, but it's going to be easier or better with us. But we have been true ever since we started the company. When we started the company, -- we call it Salesforce.com. When we started the company, we like we we're probably going to do more than Salesforce Automation, should we pick a different name, didn't pick a different name and people have told us like change your name. But we had Salesforce Automation and then you have customer service with [ Siebel ] or whatever, like, great, we will integrate. And so we always want to where they are and whatever they're doing. But we'll still pitch in these customers, the integrated platform and just all from us, it's going to be easier and probably cheaper for you long term and just cost to maintain and run.

Okay. Agentforce has been out there now for maybe 18 months. And what have you all learned in terms of adoption, sort of removing the friction, what if companies that are seeing real success with it done correctly? And how do you sort of expand that out to the rest of the year....

Who many times the people use the word for deployed engineer exactly...

Plenty.

Plenty. It's a new term and a number of other companies kind of coined that phrase. I think one of the things we learned which is kind of obvious is agentic AI is nondeterministic which we know. And so but you don't want in your call center, like it could do multiple things. We can't tell you exactly what it will do, but we'd like it to -- that's not a great answer is like how is my portfolio doing? Like do you want to give the right answer. You wanted to have the right context.

So what we found is being in the customer, and it's no longer about being in the customer in the sales process and saying, here's a demonstration of what we can do for you is we're like, why don't we build it with you. We have agentic coding now, we can mitigate the entire platform really, really fast. And we want to show you how it's working. And then we want to work with you to make it successful. So that's one thing.

Another thing is that determinism a non-determinism, in the harness of agent force, we started out just saying, well, the models are going to keep getting better. And so when I say do these 10 things in this order, that's great. It will do that. It turns out 9 times out of 10, it does. I want it at 10 times out a 10. And a lot of companies are doing this, we've pulled out some of what is really deterministic logic, which is workflow basically. And we built Agent Script, which is essentially a way and an ICI to basically script out what do you want the agent to do like coming into the website, ask them who they are or have a pilot claim for insurance is a series of steps you need to do. And in each of those steps, some of those could be nondeterministic AI through a LLM, something that kind of interaction, so mixing the 2 together. And that's been really successful is -- and it's actually faster and cheaper because you're not hitting tokens to do some of those things that you really don't need to model for.

And so what we found is, these things are brilliant brains, but you don't use them for everything. And I think what we first did is like, well, great, let's just have to do everything and turns out they're not graded everything.

Yes. And you all have obviously invested in Anthropic and a number of these native AI companies.

Yes. Yes.

Yes. So that's good was a good one.

That's a good one. Yes. Well, we like to sell [ John Somers ], but not enough. You didn't see where it was going enough.

Always too little after...

Too little, too little, yes.

but one of the questions you bring up around this sort of harness and orchestration concept is that where the value has to accrue longer term for companies that want to participate in this genic world, meaning to your point, the base level of intelligence for models will continue to get better over time. So when you think about how you differentiate, how you deliver value to customers, does it need to be your sort of ability to take that brain and then deliver sort of customer value on top of it? And is that -- do you -- and I guess the second part of the question would be like, is that durable? Meaning is that the delta between what of intelligent models, the intelligence, again, will keep get better. Is it durable? Is that sort of value-add at the orchestration level, durable.

Well, I don't think it's just orchestration. It is the -- like everything we're doing is the Headless. Because we're not building the models. We're using multiple models, mixing them for the right use cases, some for performance for cost. And when we say Headless like agent force, the entire ad bores, you can call it, harness because it's basically using these models to do customer service to do sales. We've got orchestration in there. We have telemetry for monitoring. We've got e-vals or testing the output of it can then get used to update the whole configuration the prompts and everything. And so that's hugely defensible. We've always been a CRM company. That's our wire ticker CRM. We will stay in that lane. And we're not trying to be a multipurpose like -- just like use us for any sort of AI. We're going to be CRM enhanced with AI autonomous.

And yes, I do think that's defensible. And we can also take 27 years of our customer base, the implementations, the business logic, the metadata, all of that's already out there. And they're asking us our massive Salesforce are like, hey, take us to the future because we have those trusted relationships. So I think that's also a huge advantage we have. And then we're taking them there. And we keep using these better and better models, but the models don't have the context. And they don't have the context that is secure, like we don't put all the data in the model. It doesn't have the exact right context for the question because if you put too much data to the model, it has a hard time or you spend a ton of money or both. And so all of that I think is kind of a differentiated...

And you mentioned data, obviously, bought Informatica, you had data cloud before that. How important has that been for you all to build a data platform in the back to complement Agentforce?

I mean we can call it context now because it's a cool word. Yes, we built our Data Cloud, which is really 2 things. One is the data platform for collecting data, but also a data activation platform that connects all the other data platforms out there. MuleSoft for API management, Informatica has been an incredible acquisition. It's exceeded our expectations in the first full quarter, yes. So it's been great for the business.

But I think we have too many brands right now, people know these brands, so it's fine. But we were doing customer mastering that's very important. We weren't doing product mastering. So our customers, our financial instrument would be a product, or a car from for whoever. Informatica has an amazing MDM solution for things like product mastering. And if you're an agent and you want to talk to get the right context, you want to get the right context on I'm going to the car website, and I want to buy a car, which car, you want the context, all the contacts for that product. And so mastering that is super important.

And our vision is not that all the data is there is going to take to make that work. It's a logical semantic onto logic, maybe is better where these days layer that combines the metadata of the history of sales force with metadata from Informatica, it hears all these other data sources with metadata from MuleSoft, of -- here's all these other API-connected data sources with Tableau, which is a semantic layer to understand what is the semantic meaning of all this data. All of that comes together and gives us that rich context layer that AI can then use. So it's a huge event. I'm so happy we're able to get Informatica.

Yes. Is connecting data to the agents still the biggest challenge for a lot of your customers in terms of sort of the promise, and the reality right now.

It's not connecting the data, it's the AI shows them where the data is not clean, it's not right. They haven't mastered it. we even found that when we -- I think we were perfect, but [indiscernible], when we stood up our help.salesforce.com Agentforce agent's, they started showing us where -- in our data sources wasn't it wasn't clean, was it quite right. So we had to go and fix that. And we had all the tools, obviously, with our products to do that. And so getting your data right is definitely that first step for any success with AI.

Any questions? I have a bunch more, I'll open it up. Okay. I'll keep going. I think the next one -- trying to think next one. Verticalization for you all and bringing sort of more -- it seems to me like in an AI world, the ability to bring an agent that not only understands the sort of domain and in terms of being a salesperson understands the context and then actually maybe even the nomenclature that goes into a different industry something that might become more valuable over time. I think through AppExchange, you all let some of your -- like Aviva went out and sort of originally did that in pharma.

How do you think about that going forward for you all because I could see having sales agents that are tuned for retail might be different than insurance that might be different from financial services. So I know [ David Schmaier ] spent a lot of time on this topic, and I talk to him a lot about this topic.

I mean shootout to [indiscernible] earlier, former co-CEO he really started the motion to go industry vertical, which originally, our sales engineers would just go and say, sure, I can take Salesforce and I'll just configure it for banking, retail banking or investment banking. But then we realize it's more than just the data model. And so like when you -- everybody is talking about vibe coding your CRM is like, yes, you can create a data model, but it's far more than that. And so I think we have a huge advantage as you go deeper into our product line and you look at our industry verticals, we have a lot of industry vertical business processes built out. We are building out industry vertical Agentforce agent's. And Agentforce skills and topics that you can use in your industry that understand an insurance claim, understand, I know your customer motion in banking, understand like I'm trying to think of other examples, but just to understand all of those.

And instead of handing you a horizontal here you go, it's a toolkit, go at it, we can hit out of the box, and it keeps getting better. And we're exploring with our research group. How do we -- how could we -- might we fine-tune some smaller models that are industry-specific that really underban the business process that industry to make them even smarter.

And that sort of relates to my next question, I think I know the answer will be. But I expect you all believe that this is going to be a multi-model world where you're going to be using the right model for the right action for the -- in the right, again, context. Is that happening already underneath Agentforce, meaning if someone ask a fairly simple question. You don't necessarily need a frontier model, or you might just want an open source model or some -- to your point, a small language model. Is that already going on? And how, I guess, instantaneous is that when you put in a prompt as Agentforce is smart enough to know the context of the question, so I can go to the right model, get the right answer? Or is that still a little bit...

Exactly like that. It's more like the core reasoning loop, the large foundation models are really useful like to reason what you want. To then voice has its own models to do checking on ethics or violations that could be a simpler model. Just understanding the question of what did they -- were they asking and parsing it out in the right way, can be a smaller model. And so we're -- the first step is not like a cost optimization. It's like let's choose the right model for the use case because often, it's a performance thing like I don't need to run through 1 trillion parameter model due to the simple use case and by the way, it can be expensive and it's going to take too long.

And so quality is the first step, but then performance and cost will be the next to you. And so we're mixing models all over. And we can do it at run time. We can mix and match. I think we will head in the future, we will look at should we have fine-tuned models per customer for some of those use cases that maybe we're dynamically updating the models from each customer, like we wouldn't mix the data. So that's another idea.

And finally, like we're always looking at, well, what's the next frontier model that what can it give me? And will the next Anthropic model or OpenAI, 2 biggest ones, but -- we also look at companies like Mistral that we're invested in and Cohere and other model companies who look at what do they have. We've shied away from the Chinese model for various reasons, a lot of which are -- we sell a lot to the U.S. government.

Right. Maybe you could help me with the question I get a lot, which is there's obviously going to be some workflows that are deterministic, meaning if you have a policy around CPQ, you can't just model come up with sort of guest. It can't be -- how does that get integrated? You mentioned maybe it's the agent script point you made earlier about like how do you start mixing in the benefits of both probabilistic models, but also within sort of the parameters of having deterministic outcomes to some degree. You can't have salespeople being like, all right, like close enough on discount...

Yes. I think one of the best examples is a company called Regrello that we launched, which I thing is called Agentforce Operations?

Operations.

Yes. We never changed the names of our products but it at design time uses AI heavily. So trying to understand the business process of a corporation that's not written down. It's like, well, oh, you want to give a discount on professional services. That's actually an internal example where -- we wanted to -- for a customer, I want to give them highly discounted professional services in the deal for the implementation.

Oh, well, to do that, you need approval from these 3 humans. You need to go on these 4 systems. And it can look at all the data you could draw a diagram, you can parse it or you can look at some of the e-mails that are going around. And it's using AI to understand, well, what is the real human business process, but then it takes that and it turns it into workflow because at run time, it doesn't need to be the AI running that process. It's like, first, I'm going to ask Kirk for an approval to be an e-mail and then when he really says, yes, but I'm going to go to this person to make sure the system -- and it's obvious what it is, but figuring that out, we acquired in an agenetic process.

And so I think more and more, you're going to see that. And so like companies like Dell are like, well, we're saving a ton of money. We used to call it supply chain was the first here we use because they use it in their supply chain area, but it was just simplifying their internal business processes significantly.

We obviously talk a lot about agentic. And I feel like we're sometimes in a little bit of a bubble when we talk about this in the industry. When you go out and talk to CIOs or you're talking to some of your bigger partners, I mean, how early are we? I feel like everybody wants the agentic enterprise tomorrow, but when you go out and talk to customers.

I think we're really early. I mean, we're still super early. Right now, the hot area that's getting automated, it's customer service. That's where you see a lot of little start-ups. That's where we're playing. And then in collaboration with Slack and Slackbot and see Claude Cowork as an example, obviously, coding is a huge area. But yes, I mean those are the areas that we see right now.

Okay. And any industry you think that's farther ahead, the ones that are more regulated, seem to be obvious, that will take a little while longer in certain functions.

I think it's more like the CIO. It's more of the leadership of the company, are they leaning in or not? I was just in France. I was meeting with the [ Adecco ] which is a big recruiter. And they're going all in, they source temporary labor contractors to corporations of all sizes. And I went out to one of their recruiting offices because I wanted to see our software and use. And so they were using Einstein for sales. So that's machine learning. So just like just help me understand the score some leads and score the this candidate and is this a good candidate, matched this candidate with the right thing. So that's machine learning. Then it was using Agentforce to that outbound to have an interaction with a candidate. But it was because Pierre Matuchet, the CIO is an amazing CIO, and he's forward leaning and he's going all in, and he's figuring it out.

And so I think it's -- and then it's about like are you taking the right problem to solve and there's a lot of DIY out there that some has worked a lot has failed. I mean that's selfishly saying, let us help you. So that's another thing we're seeing. But it's still super early. And I think with AI, the demonstrations are so compelling, we think everyone is doing it, and we all have this fomo, well, I got to do it. And that's why a year ago, every CEO said, everybody do AI. And everybody bought various tools and do stuff. And now we're seeing more consolidation and more use case by use case success.

One thing I forgot to ask when we're talking a Headless earlier in the conversation is, it seemed to be that Headless in a market that's moving this fast lets the customer understand that they have optionality with you, meaning you're not boxing them in, and I'd imagine at a time where a lot of CIOs frankly, aren't sure in which way they might want to go 2 or 3 years from now that's actually a benefit, meaning I can count on you all to be flexible with me because every organization is going to have to be somewhat flexible in an AI world.

It's resounded incredibly well, and we just want to meet people where they are and where they are is moving. And we have built user experiences for 27 years. We think -- and you can customize them, but we think this is the first version that makes the most sense for you for sales, service. And maybe the future is not that at all and how I interact with enterprise solutions is going to be personalized just for me. And maybe it's not me, it's my agent or agents. I mean, you look at how people are coding, they're managers of agents now, like, rate the spec and test it and do.

I think every job function will move in that direction. And we want to meet the customer where they are. So what surface do you want to do that? And what user experience do you want? If you want to vibe code a new UI for part of Salesforce, go for it. And you can use parts of what you've already configured and you can build your own. If that's valuable to you, great. If you want to use it and have it surface in these other tools, great. if you want to be multiplayer and have multiple [indiscernible] working together, we still think Slack is the best. If you want to use Teams, many people have Teams, obviously, I think Slack is way better. we will help you use Teams as a service. And I think the world is moving so fast, we can't predict where it's going to be -- we all have to be super flexible and super fast and move with the same pace.

You've obviously been at Salesforce since the beginning. Are you pleased with the agility. I mean it's a big company. So to move -- I think there's sometimes a view of like inhibitor dilemma or those kind of things with companies. But do you feel good about the level of velocity that's going on...

I credit Marc Benioff. I think he is an incredible entrepreneur. And he's like, he's talked about a inhibitor dilemma, we have to go rethink how we're organized a -- should we when we think about forward deployed engineers, we have sales engineers. Well, what's the different -- if we're not building demos, how should we think about that? Should we deploy them more out into where the business is happening. And so Headless, yes, let's go all in on it. And so I'm very pleased with the rate of change that we're driving. We have an internal process called the V2MOM that helps us stay a line, and we just keep rewriting it because it keeps changing.

But that's how tone from the top, change, try some things, don't be afraid to fail, and that's coming from Marc, and then use that V2MOM process to say like we're changing now. Now everybody we're at, here's how we need to align. But it's never perfect. I just talked to some of the leadership in our technology and product organization, like they're asking how do we take more chances and do more. So we've got to keep hammering on it.

Yes. You're obviously very in the weeds on all the tech. And just out of curiosity, what's the sort of idiosyncratic thing that you guys have broken through on more recently that perhaps only you would find it interesting, but I'm kind of curious what you're spending time on, maybe that's in the bells of the technology, whether it's data, governance model.

A lot of it has been for me personally has been in the Slack business unit. And the breakthroughs I'm seeing is, how do you think about multiplayer Claude Code where multiple people are working together with AI to build something. And that could also be Cowork, or it could be Codex or it could be linear.

So it's a breakthrough of thinking about, well, Slack as the channel-based experience you're doing work together. And it's all human based, and we're bringing agents in and what if that agent is building code or writing an S-1 to go for it. Maybe Anthropic, using Anthropic to rates on that would be interesting. But how do we do that together? And so what is that experience? What's the identity of the agent, who like -- and how do you bring it all in together. And so it's not -- maybe it's not the sexiest answer of like, oh, we've figured out this identic loop for that. But it's really more of the user experience. And I think change happens at the user experience layer.

When Steve Jobs launched the iPhone is like, well, the world has changed because of the experience, the battery didn't last day, the outport may be the best, but it was the experience. And so if you think of the company, we're really leaning in harder on what is the experience of the future, and we're trying a lot of ideas, in Slack how -- what is the experience of many people working together with AI.

And is that pretty much the operating environment at Salesforce now? It's everybody in Slack? Is that...

100%. We all in Slack. We're all using Slackbot. I mean Slackbot as an agentic agent helping people, the adoption rate is the fastest I have ever seen of any feature we built. And it's helping everybody get their jobs done, and it's phenomenal. And so everybody is like, we now have just we had a big -- we have lots of meetings -- we just had a big meeting in Las Vegas -- Los Angeles with our top 500 people. And we launched internally Tableau Analytics in Slack, but implemented for every -- all of our regions and all of our sellers, so they can run their business like instead of going into Tableau or into Salesforce or something they built themselves pulling data out and putting it in Excel got for a bit. They're just living in Slack, running their book of business and what's my pipe for the quarter? What are my top deals, what I have closed so far, who has trouble? That's all happening in Slack with Tableau tied to all the data.

And it's a great use case to go to sell because they could say like, hopefully, they're not showing everything, but Kirk, let me show you how I'm running my business at Salesforce, and here it is. And it's a really compelling way to tell.

Yes. That was my next question, actually, which was, it was a reference selling, software choice but reference, anything in the [indiscernible] is reference selling.

I mean certainly selling Salesforce automation, if you're a salesperson [indiscernible] use our tools. It's very easy.

And do you think that Slack but what you showed to customers kind of changes their perception perhaps about there -- where you can go with your technology?

Yes. [indiscernible] the team shop where they're like that. We don't need another chat tool. And we're like, yes, but can I do this? And they're like, oh, wow, and it's tied to Salesforce. And we have Salesforce channels and Slack, all the data is there and Tableau Analytics, all my work. Everything is there. It changes how they think. But we have more work to do there.

Anything in particular you think you have more work to do?

I think we have more work to do on both enabling with our teams of like telling that story and coexist with Microsoft Teams. Like in Slack, you can now join Teams Meetings, It can -- you can MCP to the team's data Black Box could use that data if you're working -- also working in Teams. But people -- most customers I talk to, they're not working in Teams. They're using Teams for video and they're using Teams for direct messaging, which is a little bit of work, but they're not really working together deeply on something. And so -- but we have more work to do on and getting some IP out there.

Just because I think it gets off a little bit. I mean, I think Marc talked about Slack getting the $10 billion at some point in time. What's the -- is the monetization sort of thought process behind Slack changed at all because of this?

Yes. Yes. So the monetization is still very much license based. We move you up in the additions. And so you want more access and more Slackbot, you move up in addition. So that's a typical motion really successful. It's one of our best executing business units. But we also see opportunities for some additional usage-based pricing to come out, which we haven't been out yet, but we're talking about working on, everyone wants their Slack corpus of data. All our partners want access to the APIs, the [indiscernible] products because now with agentic AI the intelligence you can derive from all of the unstructured data in Slack and the messages, the files, all the collaboration is a huge asset. And that's what you can see in Slackbot, but maybe one of these some other tools. And so we can monetize that as well the access to all that.

Yes, pretty amazing amount of context from a business is in Slack for a lot of them.

It's shocking, yes.

I got one more, unless anybody else has a question. All right. All right. Kind of an open-ended one, sort of maybe a softball to some degree. But yes, what's -- if you guys win in agentic enterprise, right, what would you view as success over the next couple of years? Like, obviously, adoption, you mentioned it earlier. Is it Slack becoming -- from a lot of your customers becoming the operating system for them is that what are the I don't know, KPIs to some degree that you're keeping an eye on to know that you're right track around agentic...

I would like to see Slack become the interface for getting work done for sure. And I think we are well on our way to that. I think success is also that you can clearly see how the enterprise has changed with a combination of human workers and digital workers and that's reflected in our solutions, making that possible, but it's also reflected in our numbers where you're seeing like, okay, that's amazing. This company is so much more productive because 1/3 or half of the workforce is digital labor and AR investors can see and that value is now seen in these mix of license and consumption-based revenue. And I think -- we're still on the evolution to consumption-based revenue. You see it. We've had it. Marketing Cloud has had for years. Agentforce is doing really, really well. Data Cloud is doing, you know.

But I want to see that like really clear in the market, and it's not just our pricing and our revenue there, but what has it done to the customer base? Like how does that show up where, wow, your call centers have the size and those people are now doing other higher-level jobs, and your customers are happy. And by the way, it's also a revenue-generating center because everything is blending now. It's like this was true before with the agents that they don't care if I'm a service agent, but I could also tell you like, I could upsell you on something. So that's kind of the future I see.

Agentic work units, is a something you are keeping an eye on.

Yes, it's -- we really were trying to move away from tokens. I mean, because tokens are not a great measurement of did something really get done...

Not really, token has value.

This is how much we've paid model providers and helps with their like it's kind of like leaderboards for vibe coding and what's your -- are you token maxing and who's using the most tokens to vibe code. And if you read about that, people say, well, that's a terrible metric because people are just going to try to use the most tokens and it's not really a right metric for output. And so AWS is definitely the right metric. We're trying to tie Slack to that metric as well as that drives a lot of agentic work units. And then from agentic work units to outcome also.

Great. Yes, we're right at time. Parker, thanks very much for being with us.

Thanks. Really appreciate it. Thank you for having me.

Thanks a lot. Thanks, everybody.

Thanks very much for coming. I was looking forward for this session today because Miguel and I had a terrific discussion after the company reported numbers, he's convinced I'm wrong with my rating, and he's going to prove me wrong. And I'm very open for the challenge.

So as you know, I'm new to cover software being covering cybersecurity for 20 years and networking for 30 years. And I always say that lately, I'm the garbage person at Bank of America, garbage not from a person of quality, meaning everything -- every time someone leaves, they tell me, okay, you cover it. So I cover everything else basically. That's going to change soon, but software is my focus, networking is my focus.

And I want to start with a few things before we start -- I want to start with outlining the background for the discussion. We have 30 minutes to discuss and we're going to talk strategy. We're not going to talk about the quarter. I don't care about the quarter. I talk about -- I care about where is this company positioned for the next 10 years.

Salesforce today is a much bigger company than it was 7 years ago. And the growth decelerated. The growth in the last quarter with professional services, 7%, 7.1%, without 7.7%, it used to be 30%. It used to be 20%. And that's before AI. We're not talking about AI. AI might be a tailwind, might be a headwind, right? So what I want to understand is the reasons for the deceleration, the risks that we see in AI, the opportunities that we see in AI, and how the company is positioned to basically address the opportunities because the whole discussion of this quarter, by the way, for me, it was the first time I do a quarter call for Salesforce, and the whole discussion was about AI. The whole 1.5 hours was about AI. So I want to understand AI, and that's the purpose of our discussion.

So with that, Miguel, thank you very much for coming, and thank you very much for taking the time to educate me because I need this kind of education.

And I want to start with big strategic positioning. No, I'm not giving you kind of direction with my question. I want to ask you, what do you want to highlight when it comes to the opportunity to accelerate growth over the next few years? What are the initiatives that you think will drive higher growth in the future.

First of all, thank you for having me. I will actually very much looking forward to these conversation, we sort of [indiscernible] the eyes a little bit. Can you hear me? The mic. It doesn't seem to be, if it's working?

No. All done. He's fixing it.

Feel is not working very well. Can you hear me at the back?

He's been something new.

Hello, hello. Okay. So listen, so first of all, I think since earnings is -- I've done like 10 analyst meetings. This is the one conference that I wanted to come to. You have the analysts that I wanted to meet you face to face. I've talked a lot about you. You stated about the company. I actually was very curious to understand when I saw your price target. I mean, I'm the Head of Sales and the Chief Revenue Officer, but I also started finance. I started financial engineering, financial management at MIT. I understand valuations and I was curious to get to know you. And then when I read -- I like. So I have a lot of respect for you. By the way, you are very smart. A lot great experience. You have listen, you haven't eaten a lot [indiscernible] but listen, first of all [indiscernible].

Of course, we have -- we bought a company last year that was like with the inorganic growth and we grew 13% revenue, 14% CRPO, we generated $6.7 billion of cash flow, not bad in 1 quarter. I mean there are companies that you like a lot that in 1 year, they don't generate that cash flow. There are companies that publish results the same day amazing results. We like that company. I think there are a lot of people here in San Francisco. And the whole revenue for the whole year is not even the cash flows of Salesforce. But let me...

But I'll stop it for a second because it's a conversation. In the last 2 weeks, we had a rally and the rally in software you are lagging behind. Your stock went up 3%, the others went up 30%. So something in your message doesn't get across to investors. And the fact is it's not us, that's the stock behavior. So investors are concerned. And the question is, where are they wrong? Meaning, what are the growth opportunities that you see in front of you that could prove...

Let me answer the question, but I also need to address some of the inaccuracies. So we the store went up 8.5% on Friday, it went 9.7% yesterday Monday, basically the same as they [indiscernible] or ServiceNow, et cetera. So I don't know where you get the 2%...

It was since the beginning of the year. So it went down, in the last the last role, right?

So listen, we -- our business is for 2 or 3 years, nobody really knew whether Gen AI first, then AI than Agentic was going to be a tailwind or a headwind for enterprise over companies. The world was divided. We -- I joined the company in 2011, they left for 3 years and then came back 3 years ago to be the President and Section 16 Officer and Head of revenue. And I mean, we were, I would say, cautiously paranoid about what could be happening. Last year, we saw it very clear. AI is a massive win for our business. We were competing in a crowded market where we are the absolute leader in a category, SaaS CRM. We've been doing that for 27 years. And we have amazing growth when we are small. But then when you -- as you grow, as you grow, as you grow, the growth starts being a bit more difficult and it's a multi-hundred billion TAM market.

All of the sudden AI is doing several things for us. So first of all, everybody loves AI. Everybody loves the AI labs. The intelligence utility is incredible. But in the enterprise, to convert that intelligence into proactive work, there are a lot of things that need to happen. You need to have the deterministic workflow. You need to have the right context for the data, you need to have the compliance, the government, the permission sharing. That [indiscernible] infrastructure that we built over 27 years is the big difference between beautiful AI for the consumer and productive AI for the enterprise.

So we are very well positioned. AI is making our products easier to implement, easier to use, easier to enrich and put data into and easier to consume. So it is making our different clouds more valuable for our customers. And then AI is giving us -- is opening a door of a multi-trillion, multitrillion TAM market, which is the digital labor. So today, we had the SaaS market, people continue, by the way, the number of seats, the number of salespeople, the number of service people still continue to grow. We can later talk about what happens if it start declining. We can talk about that because for some companies, in some industries, at some point, the number of seats may decline, okay? But forget about that for 1 second.

In addition to that, which is our traditional business that was decelerating, all of a sudden, you have an opportunity to monetize AI in different ways. First of all, the existing sets, we are upgrading them to our higher-end SKUs, which we call it Agentforce One Edition, Agentforce for Facility, Agentforce for Service. Every time we do that, on average, we do anywhere between 60% and 80% uplift. That's pretty big. And customers are very happily paying because now they have access to all the agentic capabilities of our product.

The second way that we monetize AI is because AI has made our software easy to implement, easy to enrich and easy to consume through conversation, then there are more seats available to us. If you look at our last earnings, 7 of the top 10 deals were customers that found new people, new humans to use our licenses or more licenses. It's the opposite. People think that AI is going to reduce licenses, so far, it's increase in licenses.

And then the third way to monetize, which is probably is 50% of the monetization that we do is what we call Flex Credits. So we basically, for customer-facing use cases of agents, those agents need fuel. So we charge them with fuel. We call them Flex Credits. And we put enough in the time so that they get going. Sometimes they buy for the next 3 years, sometimes they buy for the first 6 months. And then customers come and refill the tank. So we are seeing an explosion of Agentforce, which is one of the main drivers. Our core clouds now are better, so they are growing also, and they're growing healthily.

Then you have products like Slack that is exploding. And then we run a very diversified portfolio of products. And every quarter, there are some products and geographies and industries that they don't perform that well. But when you look at our numbers and the most important thing is, the [indiscernible] is net new bookings. So you sell new bookings, and then the net is because there are customers that are treated. Our attrition levels are decelerating and our bookings level are increasing. The difference, which is the [indiscernible] is accelerating. And we saw that last year after a great Q2, the best Q3 ever, the best Q4 ever, the largest Q1 ever that we just did, then the net NAV is accelerating significantly and is growing more than the AUV. And when the lines crossed and they crossed last year, the cross in H2 last year, they continue to surpass the net NAV growth continues to surpass the AUV growth in H1. That is when AOV accelerates organically. And that's what happened.

That's why we committed a month ago, with less visibility, but with a lot of conviction, we committed that the revenue was going to be accelerated in H2, and it's going to reaccelerate. In parallel, Informatica, the revenue on a on an organic basis is reaccelerating significantly. And then we are many other places that we can monetize, and we can talk about that later.

What is the core value of Agentforce? Meaning what do you bring with AI to customers. And you touched on it at the beginning. Why would they use your offering versus going and trying to develop it on their own. And I'm just talking generically.

Yes. Listen, Probably, I remember when we started, there was a website. I don't use it anymore, but remember, there were like 20 agentic platforms. And then 6 months later, this was 2 years ago, there were 150. Probably today, there is 10,000 agentic platform. There are 10,000 ways, there was 1,000 companies that offer a platform that you can build agents, okay? So why would you build agents with Agentforce and not with all the platforms.

So first of all, Agentforce has a number of advantages. Think of in Agentforce like an opinionated hardness or commercial use cases. So where you have direct access to the context of your CRM because it's surfaced through that harness. You have access to the hundreds, sometimes thousands of flows that customers have developed in their orgs, because when a customer implements Salesforce, what they do is they qualify their standard operating procedures in our software, and they build flows. We have a full -- they build workflows and all the workflows, it's like a library of workflow. They're available in their org. Agentforce has native access to those flows. Agentforce can easily hand over to humans back and forth. Basically, that's one of the biggest differentiators, customer, you cannot do that with other platforms. Agentforce is embedded. Agentforce allows you to choose whatever LLM you want to be on the back end. If you use -- and then Agentforce has many other advantages, and has 29,000 customers, which means that nobody in the world has the amount of customers that we have, which means that we've learned.

Our Agentforce for today, the platform, is 1,000x better than it was 2.5 years ago when we launched it. The same way that some of our competitors have increased improved. But when you have 20,000 customers across many industries, geographies, you learn a lot. By the way, some big CEOs and big personalities in the industry, they would tell you that, that agentic layer is going to be commoditized, okay? The good news is, be my guest, build agents on another platform. I want more agents. Agents are users that use platforms at a high scale. We started saying the world of AI is humans and agents working together. We changed our tag line in the company. now is humans, agents, leveraging platforms working together.

And the beautiful thing is, if you don't build your identic layer or your specific agents typically, companies use several agentic layers to build different agenting use cases in the company. That's okay. Because if you want to attach to customers, you will most likely go through our platforms. And that's other big thing that is happening. This is the one more thing like Steve Jobs. That is our Headless 360 strategy that we just launched. Every part.

Okay. So then you can ask a question.

Go ahead here. Yes. Because for me, probably the most exciting thing that we have and it's on top of what we guided.

Why is Headless different than the regular agentic opportunity? Maybe for the basics for those that don't know the company and don't know the space of Headless, why is headless different? Why is it growing? You spoke about it also on the call on the conference call. Why is headless growing the TAM.

Yes. So growing Headless can significantly or will significantly expand the TAM. We are still not factoring any of that in our H2 revenue reacceleration or in the profitable growth [indiscernible] $63 billion and Rule of 50, which is good news for investors okay? So what is different? It's a fundamental difference, before our software, by the way, most enterprise over where a full stack with the database, the workflows and the permission in everything and then we put the identic layer on top. So customers pretty much if they wanted to build agent workflows on Service Cloud, they need to do it reinforced.

Now we decouple all those layers, and we've exposed everything and we wrap them with MCP servers. So you can access the data. You can access the workflows, you can access even Agentforce, you can access lag through MCP service, which means that anyone, anyone, anywhere in any surface. I mean, many of you may be working on Slack great. You can access our [indiscernible]. Many of you may use Cowork, okay? It's very popular now. I use Cowork myself. Well, you use codec or you use another platform. So now you can use our software. In fact, I like to be controversial like you. And people are saying, okay, but people are going to vibe code this CRM. And we've been saying for a long time, that's not going to happen. People are not going to vibe code this CRM, because then they have to buy operate them and then maintain them. And now I say the opposite. I think people are going to vibe code this CRM. That's the future.

They're going to [indiscernible] because they want, those components that are already secured, they are trusted that the uptime is the right thing that you can respect permissioning, et cetera. So it's humongous. It's so big that we are afraid of we need to -- we are very cautiously we're talking and working with our biggest customers, our biggest partners to find the right monetization strategy because our #1 value remains customer success. We don't want to introduce -- we don't want to either say like our friends from Germany, oh, no, nobody access our platform or other companies that try to really take advantage of the opportunity. We just want to do something that is fair.

If you are using our platform in a way that wasn't meant to be because a human being used to the platform maybe engage with the platform. I mean, if you're in a call center, maybe 50 times a day, if you are a salesperson, maybe is 10 times a week, if all of the sudden, there is an agent doing that on behalf of a person or on behalf of an organization and a success in the platform 1,000x more. Well, first of all, our cost to serve, I mean, we'll go through the roof, so we need to monetize that, and we're working again, and every conversation we're having is incurrent constructive, [indiscernible] understanding, and it's going to be a huge opportunity for the future.

Got it. Agentforce grew like 20% year-over-year this quarter. It grew 50% sequentially. Where are we, two questions. Number one, how do you help customers adopt it? Meaning you spoke about flex spending. Can you talk about the program? What is flex spending and how does it help adoption?

And number two, where are we in the deployment cycle, meaning when you talk to your -- you have so many customers, 20,000 customers, when you talk to your customers, do you need to convinced you need to educate them about Agentforce? Do you see deployment? Or is it really at the beginning? I'm trying to understand where are we not from a technology point of view, but from a market readiness point to deploy Agentforce.

I think that's another piece of good news. I mean I'm very, very, very close to where the agentic action is happening. I'm talking to I'm very curious to intellectually to understand what is this new trend. And I would say that customers overall, forget about Agentforce, across the board, they are in a nascent period in the agentic transformation. We put -- we've done one thing that is very cool, working with external consultants. They are working with our expertise of 83,000 people and 150,000 customers. We've mapped what would be -- how would every customer in every industry. So we -- I think we [indiscernible] 17 different industries. And in every industry, we look at the top workflows and for each workflow, how the workflow will be identified. And then we come up with agentic use cases. And then we map them and then we create a library of agentic use cases with a lot of information what data they need, what do they do permissioning, got rails, et cetera.

So every time that I talk to a customer, I put a slide with anywhere between 120 and 180 use cases of agents. It's overwhelming. I call the eye chart across the workflow. They understand the workflow. I then pick up a few examples on how the process works today and how they would work with agents. But then I tell them, listen, I know the company is not ready to do that because companies need to be ready. You need to have the change management, but also you need to have the data. You need to have the legal checks and balances on any identic use cases. And I always position what we call the hero agents, the 5, 10 use cases that are powerful that I know are easy to implement.

And customers today, they are probably in the first 2, 3, 4 initial use cases of a at least of 100 and 120. So they're in very early stage. Every company, the first thing that they realize is the data is not ready. There is not enough and the quality is not there. is not clean. It's not connected, it's not been leveraged sometimes. And by the way, that's why we invest in Informatica. That's why we created this layer at the bottom, we call the Data 360 layer, or the Data Foundation layer. That is a combination of Informatica, [ Meso ] and Data Cloud is a $7.5 billion business growing double digit. And it's going to be a big differentiator for Salesforce because everything starts with the data.

And then if you have the right data, the right context, agents are not going to do ship. And this is where everybody starts. I think, every time that I tell the story, I get very well prepared for the earnings and go and go to the top 100 use cases of customers. And I get surprised because I have the same names that I talked about 3 months ago, oh, but they only had 1 agent, now they have 5 or they only have 1 million AWEs, now they have 10 million. It's live. It's an exponential acceleration. That's why the AWE use that at the end is a metric that you need to follow very closely. -- because it's real productive work, and I can walk you in a different context exactly at what -- how we measure AWE is because this is real work, it's not just taken.

That is exploding, we grew double the AWU quarter-on-quarter, 7x Q1 versus Q1. And it's -- we are just -- I think in 5 years will be probably 500x more consumption than we are today. It's that big, and we are solely in the game. And it's everyone.

Do customers feel comfortable with pricing? Meaning the whole industry goes from seat-based pricing to consumption-based pricing. How do you help your customers to make the transition?

I love you. Because we -- I have like 10 or 12 questions that your team has told my team that maybe you may be asking, and this is the first one in the list. So Thank you.

Perfect. Perfect. So the question list is just to -- I just want to make sure that I don't run out of questions.

Everything else is actually I was going to tell you at the beginning ask me anything, seriously. There's nothing to hide. So pricing has been fascinating, to see what's happened with pricing. We had listen, no idea. We were probably the least sophisticated company in terms surprising because we were selling by seat. That was our only metric basically on the pricing until 2 years ago. Then we launched a in Agentforce, and we launch it for with -- we thought we were very negative conversations. What is the conversation? Well, we don't know, but it sounds good. And it's -- then we defined it in the contract like the back and forth between a customer and agent over 24 hours, okay?

And then very soon we realized, okay, $2 per conversation, is that too much [indiscernible]. So fast forward 2 years, we are meeting customers where they are in their agentic journey. We are highly sophisticated. I mean what I'm saying is that more customers want predictability. So we have two ways to show predictable pricing. One is for internal use cases for human users of our licenses. We upgrade them to the genic version, which is the higher the premium version, typically with a 60% to 80% uplift and then they have unlimited access. We meter, but for them, we don't have meter. So no matter what they use, they pay the same. They like that a lot. In fact, those SKUs that are now is a multi-hundred probably as or $80 billion business, is growing 60%, okay?

And then the customers that want, okay, but I want to have customer-facing agents, because have a demand plan of agents are going to deploy 20 different customer-facing agents and they're going to be consuming credits and data, et cetera. So we build a demand plan. You're going to need all these credits, all this data. And the customer said, what if it's very successful and well, it's going to be more. Yes. But what if it's not successful, it's going to be less. Okay, I'm going to fluff. So we do the ILS -- unlimited the agentic enterprise license agreement, which essentially is unlimited, we fix the price, we still meter to give the information to the customer, but it's unlimited. And sometimes, customers win. And when customers win, we always win because then the renewal comes and then the customer is -- have deployed 10 agents out of 100. So okay, they want in the first 10. But at the end, we don't want to win or lose. We just want to have a fair business with a fair margin, and that's what customers want from us.

And it's predictable, it's beautiful. There are customers that are putting their toe in the water and they don't -- they just want to say, okay, I don't want to pay anything. I just want to pay when the agents consume perfect, as you go [indiscernible]. And there are other customers that say, Okay, I think I'm going to buy 1 million credits. And then you can use it in 2 years. Okay. So we are meeting customers. They are the last thing, particularly enough because we deploy some out-of-the-box use cases like agents or operations agents. And now we can price per business metric based on value. I'm telling you, I lot of business value. When I've done business value pricing in the past, we've made a lot of money.

And then customers maybe share business value, but then at the end, they say, no, I want something flat. But we are Today, we have like 10 ways to price Agentforce or agentic. But if the customer wants an 11 or 12, we'll do it too because we are all learning together. And at the end, I want the best for the customer.

Got it. We have 2 minutes left, and I don't want to completely dominate the meeting. Is there any question from the audience? Just raise your hand, we have a microphone, sure. Do we have a microphone for the audience?

[indiscernible]

[indiscernible] Yes. Premium, which typically means their pilots. Let me start with the end in Q1 in the top 10 deals, that represented $800 million of bookings and representing 2.5x more than the Q1 last year. In those $800 million, there were 3 deals that were pilots, that were successful pilots, and then went into production with a very large deal. They made out of 90,000 transactions, they may be the top 10. So they were good.

So what -- I think the -- what makes it work is having the data having the legal frameworks in place, legal is important. Legal needs to be part of the whole thing because there is a new thing about liability, share liabilities. It's -- now these agents are doing what humans you to do work many times with customers, which is what from a legal perspective, you need to cover that. And then the business value needs to be there. Then the reason that there is a change management that is very important. Typically, the good successful pilots are those where you are on top of the agent, and you are on a daily basis, understanding that's why an agentic layer, a powerful agentic layer is important to do the right testing, the right monitor and the right observability. You're listening to what the agents are doing and real-time changing and because otherwise, the agents drift.