Good morning, everyone. My name is Brian Peterson. I'm one of the application software analysts here at Raymond James. Very happy to have Susan Emerson with us from Salesforce. We're going to host a fireside chat. If there's any questions from the audience, feel free to make this interactive. But Susan, maybe to kick things off, there's been a lot of product and go-to-market investments at Salesforce over the last few years. Maybe talk about your role in those efforts and where you've been spending most of your time. All right. Good morning. I'm Susan Emerson. I'm based here in New York, and I've been with Salesforce for 15 years. For the last three years, I've been on the GenAI and agentic sprint as part of the AI product team that's now known as Agentforce. It's such a fast-moving and new space that one of the things that we do very consciously at Salesforce is make sure we've got, I don't know, trail guides to use, like Salesforce terminology, to help all our employees and customers unpack what we're building. So I'm essentially running an outbound product team, a team of data scientists, technical architects that are equally comfortable in the boardroom as they are facing off with an AI wizard. We invest in customer success and help people understand our roadmap, figure out what they should do for themselves, and in what order and why. So it's been a busy couple of years. So I know a lot of us were at a Dreamforce hearing more about Agentforce. I know it comes up on earnings calls. But as you think about the latest in Agentforce, kind of where are we with some of the key product announcements there? Yeah. So we're just wrapping up Dreamforce, and now we're on World Tours around the world. And the big things that we announced at Dreamforce for Agentforce are the following. So I'll just sort of list them and talk about, of all the things I could list, why I list these, because I think they're impactful to customers. One is we went GA with voice. And so for our customer-facing AI, voice might be a generational channel for many people, but it's a really important channel. So that we have that now in the ecosystem is great. But what I would say, I almost want to de-emphasize it a little bit, because one of the powerful things that we have is being really channel independent. You can build an agent once, and you can decide that this thing is going to one of your employees, or this is going on a voice channel, or this is going on WhatsApp, or this is going on SMS, and so forth and so on. So voice was one. And so that will help organizations where voice is sort of non-negotiable as a channel. The second thing is something that we're just calling hybrid reasoning. And over the last year, anyone who's been building agents has had to become a de facto prompt engineering wizard and dial into all the little fancy tricks that you use to try to tame an LLM into submission to do things 100% of the way you want to do them 100% of the time, which is not always 100% possible. So I'll leave that 100% thing behind. But what we did is we opened up our reasoning engine to something that is now both probabilistic with LLMs, orchestrating and creating an AI plan with determinism. So organizations now have this freedom to say, if then else, do it this way, and then have a reasoning engine do all the really creative parts. So that really opens up a lot of use cases and also control and also sort of diminishes in a nice way the skill set needed to build these things. Like you don't have to be on the cutting edge of red teaming, crazy prompt engineering. You can just build these things a lot more effectively. And I can give lots of examples. The one I often use in a generic setting is the following. Like let's say you've got a customer-facing autonomous agent, but you also have humans in your contact center, and you know that there are scenarios where people are going to want the human, or you desire to put the human in the loop. There's like many businesses where the human's a positive thing. It's not always go to the lowest cost channel mindset. So let's say that contact center isn't 24/7 or has different operating hours from a 24/7 digital labor chatbot. The first thing you want to do is you want to load up what the operating hours are of that contact center, because at the end of that chat session, if you're going to pass it to someone, you want to know if you're passing it to a contact center or you're passing it to a case management system that's going to queue it up. That's a great example of determinism. Look up the hours. If this, then that. And then there's many other examples where it's not just in the user experience, but it's in the control function. Like withdrawal of money, first look at blank, blank, blank before you determine if you're going to go do this. If you're pre-qualifying something for an event, first look up these things, blah, blah, blah. So that hybrid reasoning is going to be really important for every industry. The third thing is some work that we've been doing with, I'll just call them background agents. A lot of organizations are pretty familiar with things like RAG bots or chatbots, answer question with knowledge use cases, I would call them. We do a lot of process, AI process automation. With background agents, now we put this into a user canvas that is much like a spreadsheet, if you will. It's got a lot of ease of use to it. I'll just say it that way. Imagine rather than the AI use case cycle starting with an employee or a customer, but with a data signal and at scale, where we're running these background agents at scale, and then pulling humans back into the loop as appropriate. We call that feature grid. Then finally, one of my favorite features is the work that we've been doing with observability. Of course, we've been counting AI activity since we started it. How many conversations, how many users, average daily use, weekly daily use, like the counting of things. The counting of things is material if your understanding of what you're building is getting adopted, but it doesn't tell you if what you've built is good. So we've added all these eval models on top of our interactions so the builders of these AI agents can know if their agents are doing the right thing, agents meaning their digital agents, not their people agents. Is our reasoning engine finding the right topic to bring into the foreground? Is our AI agent executing the right task? Is it following instructions faithfully? And if we're doing content generation, is it of high quality? So these are just the beginning of the different evals that we have. And so whether you're regulated or non-regulated, you want these things because it's line of sight to is it working?, what you should build next?, and do you like what you've done? Those are my highlights. There's lots of. I like whatever. You always have to say, stop generating. Well, so actually, that's a good segue, actually, because I think in the investment community, a lot of folks have kind of wondered as we hear a lot about agents broadly, is there kind of a build versus buy decision for a lot of enterprises? So as you think about all of those things that you highlighted, when you've talked to customers on that build versus buy decision, how is that weighed in? What have you seen? Any help there? Sure. Build versus buy is always there. And in an inflection point like AI, of course, it's there very heavily. And so the conversations that I usually end up having are about, of course, there's build activities happening in every organization. The question is, where should you really go in anger with the Salesforce products? And how do what we have with Salesforce, how is it not just amazing for the community of users that we have with customer-facing applications or employee-facing applications, but how does that integrate with everything else you've already decided for? So I think it's more like a question of that. And if you've listened to the earnings call, we're seeing super ample evidence of this as a successful strategy with the number of Agentforce deals going up quite significantly. I think we reported 18.5K. People are re-upping the gas tank in terms of the fuel to run these agents, like 50% quarter over quarter. The amount of folks going in production, 70% quarter over quarter. So we're just seeing tremendous growth. And then I'll tell a story about one of the organizations that I met with last week in Europe about this sort of build versus buy that kind of pulls some of our product capabilities through. They're a large insurer, and their entire backbone is Salesforce. They use Salesforce for selling a new customer. They use Salesforce case management for processing every new underwriting activity and every renewal of that activity. It is end-to-end wall-to-wall Salesforce for the most material part of their business, which is underwriting risk and talking with customers. They see AI as the fuel to take friction out of every bit of the process, and the process is Salesforce. Now, on the other hand, they got a bunch of young AI studs who would love to build, but there's not enough time, money in their business model to support that. So I started showing them some of the work that we're doing with our new builder. I mentioned the hybrid reasoning. In addition to the hybrid reasoning, one of the things that we did is we made some user experience changes. So you can be the traditional accidental Salesforce admin and drop into something dead simple and build an agent, but you can flip to code and scripting and an agent graph in a beat, and when the head of the transformation department and the CIO and the head of Salesforce saw this, they're like, "This is it. This is what unlocks it because you're bringing tools that are commensurate with our AI team in conjunction with our business transformation thing, two different canvases on top of the same engine, and will allow us to go a whole lot faster without armies of people doing tool integration for years and years. Second comment, one of the largest financial institutions, one of the large ones that I work with here in the U.S., they said this to us. We do do-it-yourself to learn. We use packaged software to scale. There's a lot of learning and getting your hands dirty with all this stuff. But if you want to do this stuff at scale, you don't want to be responsible for bringing it all together, maintaining it. And Salesforce is an obvious choice for them given the operations they run with customers and employees. So I love that statement. And I guess maybe I'll bring it back to kind of budgeting and how people are looking at where the investments come from. So are there innings of AI, so to speak? People experiment, and then they scale with Salesforce. As you have these budget conversations with customers, how are those evolving? I guess the budgets come from everywhere, from just the operational budgets that people have to run their business, and they look to modernize it or rationalize it and bring more to Salesforce. There's the budgets of transformation. I mean, this is still like three years in. I would say in year three of all this AI, it's back to the boardroom again. In year one, it was the boardroom because everyone was like, do I go out of business? What does this mean for my operating model? In year two, it was like experimentation, learning, piloting. In year three again, it's like, OK, we get it. RAG bots are cool. How do we do transformation? And transformation budgets are in the office of the CEO. And then for some organizations, it's interesting. We talk about digital labor. For many people, the concept is it's vague until they see it themselves. I'll give you an example with one of the customers I work with in the recruitment space. These guys have been public in different domains. I won't use their name. At a recent conference, they came running across the room to me. Given my role in working with customers, sometimes I'm getting yelled at. I was like, oh gosh, he's coming at me in full force. What am I going to get yelled at? He's like, oh my god, it was amazing. Did you see what happened? I'm like, yeah, I saw the log files. You're processing your recruitment pipeline. He's like, yes, it's amazing. He went on to say, what's really amazing is that people are looking for jobs after hours. Like, yeah, of course they are. They're not interviewing while they're sitting at their desk. And I'm like, that's digital labor. It's 24/7, 365 when your employees are tucked into bed. And then he went on to say, we're seeing the opposite hallucinations. You know how everyone worries about hallucinations, and you plan for it, and you build around it? He said the AI agents are way more responsive to the instructions we give them. In fact, they follow them way better than the humans. And the end result of this is that we have a stronger pipeline of candidates with a higher acceptance rate. I'm like, yeah, this is digital labor. And so it's sort of one of these things. It's abstract until you see digital labor taking on real workload. I mean, our story internally at Salesforce has been the digital labor that we have handling over 80% of our inbound inquiries, which used to be human-led now. And now that's capacity that is freed up for much more interesting things. One other topic, I was in a panel last week, and I won't use their name because it was sort of a Chatham House Rules thing. They're in the health care space. And they've been using us since day one with everything from answering questions about how to prosecute claims for health care. And what they have done is they've banked a lot of money in the following way. They have reduced the number of people in their call center because they can do a lot more with digital labor. They did it all with natural attrition. No one was fired. Now these people are all earning much higher wages, and they're doing more comprehensive things. Whether you use these savings to bank into new things or you're finding new budget because you're a brand company and you have to build an immersive, dynamic, amazing next-gen experience, we haven't found budget an issue, I guess is what I would say. For people that have unlocked, that's a great concept, digital labor. What have you seen or maybe some kind of customer success stories for those that have really embraced that and have kind of in that year three evolution with Salesforce? Any customer examples you can highlight, and where are they seeing that broadly in terms of their evolution with Agentforce? Well, I just used three. Salesforce, I can use our name. And then I used the example of our customer-facing one. I used one of a health care company that now is processing claims and answering questions from members at a much greater scale. And then I used an example of a recruitment company. I'm working with a couple of financial institutions where, just based where they are physically in the world, they have an opportunity to go to adjacent companies or countries. And they don't want to do that with human labor. It's too expensive. So now that they have these different AI agents that do everything from answering questions about product and services, prosecute the KYC and customer onboarding, schedule meetings with bankers, those types of things, they're using this to aggressively expand to new markets in ways they could never conceive of before that would have required not just human labor. But when you think about this is one of the topics we see a lot right now with AI. The general statement is, with this shift in AI, the cost of intelligence approaches zero because everyone can be enabled with an AI. So what does that mean by the types of people you hire in your organization? And Ethan Mollick had a little quote out on LinkedIn about a month ago. If everyone can market a little, everyone can sell a little, everyone can code a little, what are you hiring? Are you still hiring the person that's built 20 years of experience around a certain domain thing? Are you hiring this jack of all trades? So in the example of this bank, they're hiring the jack of all trades because these people are doing onboarding, underwriting, selling, marketing. And so that's another example. Sorry, just something I could build on with that. But in terms of pricing, I know that's a debate that a lot of investors are having. You guys have evolved your pricing model a little bit. How do you kind of balance seeing the value that everybody's getting, but also maybe kind of the predictability of wanting to control costs? How have those discussions gone with customers? I'm always really critical of our pricing. And I'm really happy where we are right now. And I'll just mention a couple of ways we approach it. It's been a new category, right? And we have been experimenting over the last three years about how to do this. And what I like about where we are right now is that we've got choice. And so choice in the following ways. If you have humans in your workforce, for sure, every process is going to be lubricated with AI. So you don't want to be thinking about forecasting it and counting it. You just want to use the stuff in anger. And so for those things, we have the standard per user per month, our Salesforce buyers think for that, and a lot of bundling strategies. So we've taken the friction out of it for employees. In terms of externally facing ones, if we're doing things that are like the call deflection and the customer ones, call centers are nothing if not measured, and they usually can give you a lot of detail about the types of calls they have and the reasons they are, and so there is operating runway for that, and then we give a variety of ways for people to buy, whether it's pay as you go, pre-commit, or pre-purchase, so there's a lot of flexibility there. We've landed as the unit of measure, not token, token, token, token, because what does that mean? Does that solve anything, so we are really basing things around the concept of an action? Did the AI do something and manage a task for you and bring some automation to the foreground, so that's sort of the unit of measure, just to be transparent on that. Finally, we've been this whole idea of we don't want to forecast it. It's hard to measure. It's a new category. But we've chosen new Salesforce, and we want to cook. We've got these unlimited Agentforce enterprise license agreements. So I've never been happier where we are. And I'm always calling friction on things I don't like. And I think we're in a good place now. What about in terms of competition as it relates to AI? I feel like there's going to be a lot of things that are new. I guess, how do you think about competition in kind of an agentic world? There's always competition as long as you're in a real market. There's a lot of competition here, as you know, everything from wee little startups to hyperscalers. It doesn't matter what industry you're in or what category of anything you're in. You have to have a uniquely differentiated advantage. The things that Salesforce has going on for it are, we've got the context of everything that is customer-facing. We've got the workflow. We've got the processes that are either de facto or material through things like Lead- to-C ash, through growing that relationship, through servicing the customer. That's just not simple RAG bots. These are processes that are automated in a system of record like Salesforce. We've got the context. We've got the action. A lot of people might kind of throw a marker down and say, data. Data is gravity. Data for real is gravity. I mean, we did buy Informatica for the ability to unlock a lot of corporate data. But what I would say is almost even more important than the data for grounding things is the openness and the mindset of your architecture. And at Salesforce, with our AI suite, we are open at every level, not because we had to, because deliberately we want to. So we are open to data, whether we're acquiring data via MuleSoft, Informatica, via our Zero Data Copy. We are open to LLM choice, whether we're talking about Gemini, OpenAI, or Anthropic. We are open to things like MCP and A2A. And so if you're a buyer of these things, you don't know what the next corner is going to be. No one has the crystal ball of what amazing thing is going to happen in six months or 12 months. So you want to be future-proof with openness. And so this openness is a very compelling capability we have. And then finally, back to the old adage of everything is possible with time, money, and code, but you never have infinite amount. We've got all sorts of ways for people to go fast, whether we're talking about the skill set of who's building or the fact that we know the job to be done and the persona across sales, service, marketing, Tableau, and about 12 different industry clouds. And we jumpstart these things with out-of-the-box agents that have the context of our data model, have the context of the job to be done, have the actions behind them. And so you can start with these things and go. And because it's an open platform, on day two, if you change your mind, you fiddle with it until you're. So maybe talk about it, because there's a lot of Data Cloud stuff, the importance of Data Cloud and agentic. And how do people kind of balance what's first? Do they need to get their Data Cloud in order, and then it's a gentic? Or because you're working with customers, I'd love to understand how they're approaching that. My guidance to customers is always to start and not have data be an excuse. We know we need good data to ground these things for accurate AI. There's always a good enough pile of data somewhere to start with a use case. So we would sort of be hesitant to say, go off and do a five-year data project because you're going to miss all the benefits of AI in the short time. One of the things that I think has been we're really starting to see the market grasp this and is Data Cloud. I mean, Data Cloud is a cloud. It is a CDP, pure play for people who are creating their marketing assets where they have to harmonize customer data and market to them. But Data Cloud and the rest of our infrastructure at Salesforce is an activation substrate. We don't need to have the data in our application to advantage it with this thing we call the Zero Data Copy Network, where if people have their lovely Snowflake Lake or their GCP Lake or their Databricks Lake, we can leverage that without moving it, without rematerializing it in Salesforce. And so that message is finally really starting to take off. And I think in our earnings call, we talked about some of the process. I think it's like 32 trillion records in Data Cloud. Over half of that is with this Zero Data Copy stuff. So I'm very happy that we renamed it to the Data 360, because as soon as you say Data Cloud, it suggests we want your data in our cloud. And yes, we can do that, but we also think it's more strategic to be this activation substrate. So as we think about Data 360, then what has Informatica recently closed? What does that bring you in terms of the comprehensiveness of that portfolio? Corporate data, MDA, lineage, cataloging. It just opens up the aperture of all the ways we can ground this AI and activate it into customer and marketing and selling processes, supply chain as well. OK. And maybe just it's a little bit different from the technical side, but I know Hyperforce has been a key investment for you guys. How can you talk about some of the benefits of that? Where are you in those efforts? And is that potentially unlocking cross-sell? I mean, we've been on the Hyperforce journey for like a decade now at this point. So I kind of personally just take it for granted that we have infrastructure in region where it needs to be. To be honest, to me, I don't even see that anymore because it's just an assumed benefit of course, we have a data center in country X, Y, Z. But we're also moving to GCP, which will be really exciting for organizations that have Google as part of their infrastructure. I think one of the key narratives over the last few quarters has been some of the sales and marketing investments that you guys have been making in terms of capacity. Can you talk about where are you putting that and maybe some of the benefits in terms of growth that we should expect? Yeah. What I would say is, when people are asking me about AI, it's usually like, what's in the roadmap? What's going on? But AI has transformed Salesforce in every aspect, whether it's pricing and packaging, whether it's the way we deploy our human capital, whether it's selling human capital or these things we call forward-deployed engineers. It's been a big shift for us. And Marc sees this capacity that we need all the time. So we've been making very strategic decisions around capacity investments. And what we've been doing most recently on that to take advantage of the growth that we see is increased capacity in AI and data sellers, increased capacity in the SMB channel, increased capacity in Life Sciences with our new cloud there. And then secondly, I mentioned these forward-deployed engineers. This is a new category of software. Just because people have certificates they can go get and training and enablement programs, they can go. The investment that we were making to help people learn, adopt, and implement has been really quite significant as well with these forward-deployed engineers. And maybe just kind of lastly here, the $60 billion target you laid out, where do you guys see kind of the biggest incremental opportunity to expand with customers or add new customers? It's just like, never been a more exciting time. I was in a conference in Oslo last week, and I was listening to our country leader talk about the types of proposals that he has been generating with customers, and he's never seen anything like this in terms of the potential for innovation and growth with this kind of technology. It's way more than contact management and opportunity management for selling and way more than case management for servicing. It's real transformational stuff, and so at the highest level, what I would say I see is that the Agentforce and Salesforce, it's the same conversation now. There's no Salesforce without AI, and there's just this flywheel of benefit across the processes and user experiences with Salesforce with AI, so there's the potential for every process to be automated with AI. There is the potential for one AI use case to generate a whole bunch of other use cases, because now with observability, we see all those utterances. So we see what the people really want. We know what agent to go build next. And then with this flywheel, Salesforce is gifted, blessed with all these different clouds. So it's an opportunity to go multi-cloud. And then with the acquisitions of Informatica, just going even deeper into the operational system. So I just think it's never been more tremendous. That's great to hear. We'll end it there. Susan, thank you so much for your time. Thank you. All right. Thank you. That was great. All right.