I pulled myself together for you, Bill. I'm in good shape. Thanks for coming. Thank you for having me. Yeah. We have a full room here in part because you ended up putting up an interesting quarter. We're not going to go through with Bill the details of the financials, mostly because Bill's focus is a little bit more on strategy and, in particular, product strategy. So that's our focus today. Bill, do you want to elaborate a little bit on the role at Salesforce so people have a little context? Yeah. So thank you for having me. Like I said, it's a pleasure to be with you this morning. It's always the most popular session starting day three, first session in the morning. So it's great to see so many of you today. My role at Salesforce is I'm the EVP of Corporate Strategy, which we define as really defining the markets and landscapes that we sort of take and compete in with our software, our monetization strategy, how we think about pricing, packaging, bundling different solutions from across our portfolio, really considering new categories of expansion, total addressable market expansion for Salesforce, and as part of our product thesis, really where we make our investments over the next 12 to 18 months of time. So my team has been deep at the heart of planning around our Agentforce sort of transformation, really thinking about how does this enter into a new space around the CRM category and really front office of software, and really everything from how we price it, how we package it, how we bundle it has been the major focus for our group the last year or so. So we're definitely going to spend some time on Agentforce and take some questions from all of you. But maybe to start at a higher level, given that there's more than just Agentforce going on at Salesforce. So, you know, obviously over the last several years, five-plus years, Salesforce has acquired and pieced together a much broader suite. So maybe one high-level starting point, Bill, is what's sort of the common thread or product strategy goal? Because at first blush, somebody might look at Slack and MuleSoft and Tableau and others and not see how they fit into the whole. But what is that common thread? Yeah, I think for us, it really hit a sort of a crescendo moment just exiting the pandemic where we saw this really heavy focus towards digitization just of the workplace. If you recall, and it wasn't too long ago that we wouldn't be here in person. We'd all be at home and all digital in sort of the way that we were working. And for many businesses, especially businesses that maybe were laggards in the digital movement, it was sobering for how they had to sort of think about reopening their operations. So this remobilization of labor around the post-pandemic era was one of the greatest movements since like the end of World War II. And so what we really thought about at that time was we needed to think differently about what front office software truly looks like, how we communicate, how we sort of engage with customers, how we really analyze data. And we were generating so much data because the world becomes so much more digital that we needed to kind of process that, help people process that information and put it into places that mattered, like when you're interfacing with customers. So really since the post-pandemic era, we thought about, OK, how is work changing, which led us into a thesis of Slack from an enterprise collaboration standpoint and from not just inside the company, but really outside of your walls for digital communication. We talked about, OK, how do we think about data and data coming together in new ways? And that's really led to the birth of Data Cloud. We talked about AI and sort of this need where with more hyper-explosion of interactions, companies were just struggling to keep up with how they sort of think about interfacing with routine customer interactions like service and sales. And so that really led us to this AI movement. So everything for us has been this building and this reaffirming our strategy around communication, data, customer. And that led us to sort of the place we are. And I would say in the last five years, the company's portfolio has become incredibly impressive. We think about the analytics capabilities of Tableau, the integration capabilities of MuleSoft, now our data family expanding with our acquisition of OwnBackup. It really becomes this true assimilation of front office computing that we're really inspired by. And Bill, where do you think you are on the journey to tightly integrate all these disparate, in many cases acquired, but sometimes organically built segments? So you kind of 10 out of 10 on the integration or more like seven or eight out of 10? I think we have more room to go. I don't know if I'll give you a score on that. But we talked about this in our earnings announcement just two days ago around an initiative we internally call our More Core initiative, which is all about how we take these platforms that are best of breed in their own sense but bring them together in a core sort of foundation of metadata, of business process logic, of security access, privilege access to information. We're well on the progress there. We've been doing it for about 18 months. And I think we probably have another 12 months or so to finish the job. But I do think that has been a big thesis of reinvention is if we could take Sales Cloud and Marketing Cloud and work them together in a way, we help organizations produce revenue better. I think that really becomes truly an unfair advantage with the best of breeds that we have in every category. And there's probably a margin benefit too as you pull that off. Doesn't hurt. Okay. So let's flip to AI. So we've heard a lot from Marc on the last couple of earnings calls. He's taken up a couple of minutes on each of the few earnings calls to describe it. How about your contribution, Bill, to this opportunity sitting in front of us? I think one nuanced question that everybody has is I think we're all reasonable enough to know that despite the initial marketing hype, these things don't happen overnight. They're kind of slow wind-ups. Maybe you can talk a little bit about sort of the timeline and trajectory here, knowing realistically that agents are not going to do everything tomorrow, but it'll be a progressive improvement. Yeah, I think we still are in an era of heavy experimentation for a lot of companies. And most organizations that I deal with, and I deal with some of the largest sort of CIOs, developers in the world at those organizations, they're still sort of experimenting and really figuring out how far can we push at sort of the bounds of this technology. What I'm encouraged by, though, are the stories where you really see these early wins that are like these rays of light and truly rays of hope for customers that are struggling. And I look at a customer of ours like Wiley and the back-to-school movement for textbooks. And every year they had always struggled to sort of keep up with the demands. And so they hired big labor forces and then had to retire those labor forces sort of post those surges, if you will. And they struggled with it, struggled to train people, struggled to find people, struggled to get people to really handle the quality of interactions that their customers expected. This ray of light of AI and Agentforce enters into the fold. And now the training was just a matter of days, not weeks and months. This surge was something that immediately you put online, and they're seeing a faster resolution time than they've ever experienced before. And so while, again, we were in a period of experimentation for many organizations globally, you just see this thirst, this need, this desire to maybe think differently. And I think that's truly where most customer sentiment is, high excitement, but also a little bit of just maybe hesitation for how far, how fast, how quick can we realize the benefits. And Bill, can you talk a little bit about the projected rate of performance improvement in the agents? It's the level of complexity or let's maybe confine the question to customer support agents because that's an easy big one. What can Salesforce's agents handle today versus what perhaps can they in six months? And maybe talk a little bit about what the catalysts are. I think you guys have teased out a little bit that there could be a version two coming. When is that? Because we all just want to monitor that performance improvement. Yeah. So we've been working on service deflection, service automation for many years now. I came to Salesforce in 2017 to lead our Service Cloud business. And that was the first time we really put our innovation into chatbot technology. Raise your hand in the room if you like dealing with chatbots. No one, right? No one likes dealing with chatbots. But in chatbots today, for a lot of organizations, do about 40%, 50% deflection of a lot of sort of interactions. And they are pretty much menu-driven or contained to a set number of topics or sort of directives. What we see with Agentforce and what we saw with Wiley is you're talking really from the first experiments, 70%, 80%, 90% sort of containment, not deflection, but containment where you're resolving the issue, not sort of directing you to another piece of content that exists. So when you think about that kind of productivity gain, you're really talking about moving to an area where work truly gets completed as opposed to work gets deferred. And I think that's really what is truly exciting about the rate and pace that our customers are realizing in the early days of Agentforce. And I think that from an innovation standpoint and for where we're going to take this technology, especially in I'm just looking at my watch in two weeks' time when we announce 2.0. We just did 1.0 in late October. We're already on a pace of innovations moving very, very quickly here. It's really all about the speed, the performance, the diversity of working with different sort of languages, models, data sets that are out there, and looking at how we can put Agentforce into new places for customers. We still are conditioned that agents sort of exist in the chatbot sort of metaphor. I think what you're going to see in a couple of weeks' time is agents really become experiential all to themselves, which I think will be exciting. Bill, what's the catalyst for those agents to get a little bit smarter? We'll wait a couple of weeks to see what version two looks like. Maybe at the risk of being too simplistic, is it that the underlying LLMs that power the Atlas Reasoning Engine are improving, like you're upgrading the models? Is the performance improvement from the agents a little more related to the customer proprietary data sets that you're exposing to the models? How does the performance improvement happen? Yeah, it's turning out that the rate and pace of model enhancement and development has started to slow. And so it truly is the data, and the data that you gain access to, and the data that you can assemble. And this is why our Data Cloud strategy has become such a kind of key linchpin in the success of our AI initiatives, which is really giving more secure access to that data, feeding the model with that data, and helping the models now reason over that data. I think that is truly the biggest innovation breakthrough that Agentforce has created is the reasoning on the data that is your data, not just data that exists sort of in the public domain, but data that exists really in your corporate and private domain. Bill, do you need Data Cloud to really have a customer execute well on the agent plan? Or is it more that aggregating all this critical customer data in one place makes the rate of improvement in the agents greater, but it's not absolutely necessary? Yeah. There might be a little confusion about that. Yeah. So Agentforce is best with Data Cloud. And Data Cloud, as a sort of fundamental architecture of the Agentforce platform, really is about grounding your enterprise data. What I think a lot of confusion exists is, does that mean you need to reinvent your data lake strategy? Yeah. The answer is no. We have a great open ecosystem with the Snowflake, the Databricks, the BigQuery of the world where if you have that as your intended strategy, we can work with that data in a zero-copy way. So I think people maybe misconstrue that, OK, Data Cloud must mean an alternative strategy to your data lake. That's not the case. But Agentforce does require Data Cloud in order to ground that information. You mentioned at the onset that the most recent acquisition that Salesforce has done for about $2 billion was OWN. While we're on the topic of Data Cloud, it might be the right moment to ask this. How does OWN fit into this Data Cloud strategy, Bill? Yeah. Well, there were two acquisitions that we made. One was Own, which was a little bit more around getting access to different kind of data sets that might be enterprise data sets or secured or backed up, et cetera. But the other one was a company called Zoomin. And Zoomin was about really taking unstructured data. So if you think about our thesis of data and how we bring that into our agents, data is not just sort of the things that sit in rows and columns in the database. It's also oftentimes in your content, your communications, your websites, your other applications where you might have that information. So our thesis was, how do we take data from both the structured and unstructured way and bring them together as part of our Data Cloud strategy? OwnBackup helps us get structured data, and Zoomin helps us get unstructured data to make that come together. OK. That's helpful. Maybe we'll move the conversation away from performance improvement and the underlying architecture to pricing. So you mentioned that you also have your hands on the Agentforce pricing. So maybe you could lay out for us, how did you guys collectively get to this $2 per interaction number? What's the logic behind it? Yeah. So I think, like I said, we've been working on chatbot technology since really 2018 or so. And we've always seen how chatbots are sort of in the sense for interaction kind of model. And what we really wanted to do with Agentforce is not sort of peg this into the chatbot sort of era. This is a different kind of software, a new kind of software that's never really been created, which is really around truly human or digital labor creation, if you will. And so we took great inspiration from many of our customers who were working with us and saying, you know, today our average customer service professional is about $17 an hour. They do about two-three cases maybe an hour. And that's per cost per call for us oftentimes is $5, $10, $15 for interaction. And so what we said is, OK, what if you could hire similar talent and similar labor to do the same kind of resolution, but in the kind of less than $5 range? And they sent clear signals to us where we would hire in your labor force and stop offshoring labor because your labor force would be the cheapest one we could hire in the world. And we could also train it, recruit it, make sure it's of quality faster than ever before. So we really got inspired by this notion of labor augmentation. And I think the $2 per conversation, based on sort of our early returns, when you actually show the ROI to customers about how many calls do you handle versus with your human service center, how many do you want to handle with your AI center, the material kind of benefits are quite profound. So we've actually gone public with our ROI calculator for our customers now. You can find this on our Salesforce.com website. You can actually see it. And so every customer that's really gotten Agentforce in just has huge gains, even at that $2 per price point. OK. Bill, let's now talk a little bit about some of the alternatives to using Agentforce that Salesforce's reps are going to have to sell against, so the one that pops up most frequently when we do calls with large enterprises is, by the way, I should preface it by saying the good thing is that many of these large customers are beta testing Agentforce, so early on, it's pretty clear to us that perhaps your incumbency or early advantage on the support agent side has, at a minimum, secured you a seat at the table, that's victory number one, but they are looking at alternatives, and the alternative that pops up most frequently is to just DIY the agents. Yeah. And I know that's evidently gotten under the skin of Benioff because he talks a lot about that. But why are customers even evaluating DIY, Bill? Because if we look over the longer arc of the last 10, 15 years of tech changes, you were at the forefront of one of the last ones, basically pioneering this whole SaaS trend. That wasn't a DIY option. It went pretty immediately to third party. So why are we in a different cycle now that this is even an option? When we entered the cloud market in 1999, the number one competitor was Siebel. And then the secondary competitor to Salesforce was custom build. And so there always is this sort of sentiment around your customer interactions, your business processes, your data is quite unique to an organization. And so therefore, the first move that a lot of organizations make is, let's go start kind of on the building side. And so this is not something that has been unfamiliar to our company around the DIY movement. And it's why, really, since the inception of Salesforce, we've had a great platform that allows organizations to tailor fit CRM into their unique business needs. Your organization, for example, works differently from your competitors. And so what we need to do is make sure that the CRM applications sort of fit in that regard. Where we are in the world of AI is very similar to that right now, where a lot of organizations started their first movement, which is, OK, a lot of vendors came out and said, OK, you can build. And it's just turning out that model and model development and the quality of outputs around sort of setting the right guardrails for those models has just led to too many hallucinations or led to, example, like a Microsoft Copilot, where data and data privilege was held outside of the bounds of your applications. And so privileged information was sort of leaking across the enterprise in ways that was just not secure and not safe for companies. So we found great inspiration, I think, in this moment to go back to our roots, go back to our roots of being the platform that you can trust, being the platform that is built with guardrails and safety in mind. When we entered into the generative AI era, the first thing we did was we built a trust platform, which was really about taking toxicity out of response or taking kind of sensitive information around sort of my PII data and saying, never send that to a model. So this foundation of kind of working on trust and configuration and making it so it can be tailored, not have to be coded, was really going back to, again, what started the company in 1999. So I think this moment, again, we find ourselves in many customers are still in that, do I DIY it? For those of you that don't know what DIY means, means do it yourself. I sat next to people at Dreamforce when Marc was out on stage saying, DIY. They're like, what does that mean? Is that a different kind of AI? No, it's not that. We were literally helping our customers really rethink their strategy about going faster, but also having the quality of output, having the security of that data kind of be part of that integrity of the solution, and allowing kind of every company truly to take advantage of this much faster than they've been able to. So maybe we'll pivot to more conventional alternatives, not so much DIY, but other software incumbents that wouldn't mind a piece of this market. Bill, you're not going to have it to yourselves. I think you know that. So Microsoft at Ignite, Satya talked a lot about integrating agents with Copilot and going after that opportunity. ServiceNow is now talking about agents. Meta is talking about agents and hired one of your executives to push forward on that. We had Sarah Friar, the CFO of OpenAI, on stage the other day here at the event. And in her judgment, 2025 is the year of agents. Now, they're all probably going to go after different agent end markets, not necessarily customer support for sure. But I guess my question to you is, what's Salesforce's right to win when you've got giants chasing the same opportunity? Competition is a beautiful thing, especially with where we are right now on the sort of dawn of a new age of software, and I definitely think that every customer in the world should try as many of these platforms as possible because every one of them is going to provide a unique sort of point of view around sort of what the innovation can do. Why I feel quite strong and bullish about Salesforce's position is we believe that the world of business software and the world of sort of business process is going to be the partnership of humans and agents working together. We don't think that the world is going to be binary, where you have humans or agents. We actually think it's going to be a world where we're actually working together to solve the needs of our customers. And on one side of that, that means that the company who really understands the interactions and working with the humans really has an advantage, because you know how you've been dealing with customers for the last 20, 30 years already as part of that foundation. And so our thesis is, by augmenting human capacity with agent capacity, you have better results. And so I definitely think our place is in that sense of partnership. Great human agents, great virtual agents, all working together to serve your customers. Maybe we'll just press on that a little bit because it would seem that if a customer is going to justify a big investment in agents, maybe the justification can simply be the speed and accuracy of their responses to the customer. And that might create enough value to justify the investments in this new technology. But the ROI to the customer is more powerful when they can actually start slowing down the pace at which they're hiring contact center agents or, frankly, eliminating them. That's the big ROI. So is it not likely, Bill, that over the next several years, as your agents and others get more powerful, that we will see seat compression? And for investors worried that you might win on the Agentforce side, but don't forget, you might lose on the seat side, what would you say in response to those concerns? I think we've been forecasting a death of categories for a long time. 2013, technology called robotic process automation came out, and I remember reading about how that was going to be the death of the contact center in 2013. It didn't actually happen. What we actually saw since 2013 is the seats have actually grown around the kinds of interactions, but the kind of labor that your agents and your human workforce actually performed just looked differently. Some of it was more productive. Some of it was kind of higher empathetic work. Some of it was being able to maybe think about new skill sets into your service center that are now your sales center, so a blending of the workforce in some sense. I'm not a big believer in the death of workforces. I'm a big believer that the customers I talk to today are already having trouble hiring. They're already having trouble keeping up with the demands of their customers. Customer demands continue to rise. They have so since really the end of the pandemic, and I do think that what my customers really see is a chance to finally catch up and a chance to finally close the gap of what those experiences or customer expectations truly are telling them. Now, it's not to say that new categories of software aren't going to create maybe a displacement of different labor types, but it's also going to create new jobs. I remember when the world of social media came out, the marketing teams all freaked out about what does my job look like. It gave rise to new jobs called social media managers and influencers and a whole other categories of sort of work have sort of originated with this great time of evolution. I think that's more what's exciting in the eyes of our customers is not just displacement, but creation, and I definitely see that as a future that we're happy to help lead. So in our last maybe six minutes, Bill, maybe we pivot away from AI because, well, you might have come away from Dreamforce thinking that Agentforce is the only thing on Bill's mind. Probably there's a few other things. So as the head of product strategy, what are the other big initiatives that you're thinking through with your team? Maybe it is all Agentforce, but I'd like to ask if there's anything else. Yeah, there's a couple of themes. I talked about sort of the theme of how organizations communicate and how really breaking down silos in companies is really a big thesis that we have with the Slack business. I talked about the data movement and how we're kind of bringing new types of data, diverse data sets into the Salesforce world with zero-copy partnerships or OwnBackup or our Zoomin sort of acquisitions. So those pieces have already sort of covered. The one I haven't talked about that I'm really excited about, I'm just not excited about the form factor, is spatial computing. I think that. What is that, Bill? Spatial computing is really the blending of physical and digital worlds together. Has anyone here used the Apple Vision Pro? It's a magical device. Expensive, but a magical device. This is really about breaking new barriers, I think, for companies to go into the last mile of where their customers may fall, and so solutions like our field service team and what they're doing around the virtualization and the robotics nature of kind of helping companies bridge the gap between physical and digital is something I think is another big opportunity. We never fully realized it with IoT at the time, but I think it's sort of the full maturation of what comes next, where more data, more intelligence, that process can actually help advise companies to take advantage of that sort of computing platform, so I'm excited about that. I think that will be another sort of area of investigation for us, especially around businesses like our field service business, our MuleSoft business, et cetera, and I think that also maybe beyond that a little bit is the evolution of the business process itself. I think so many business processes were codified in a world where humans were expected to be part of every step of a journey, and now with humans and agents sort of working together, I think business process automation and business process work will also sort of evolve, so not so much business process management. What does the future of that category look like? I think that is another area of great interest for us as well. Just on that note, business process automation in terms of front office, Bill, or looking a little bit broader? Yeah, I think maybe looking broader. I think we've been for several years had this divide between front office and back office. And I think what you're sort of seeing is the rise of the middle office is creating a lot of challenges on both ends of that spectrum. So I do think that will be an area where, as the world sort of kind of immerses itself with agents and humans together, we'll rethink the business process entirely. And things like speed and scale and availability of people to be part of that process maybe becomes less of a constraint than it once was. So I think that will be an exciting area to innovate. And I'm excited to not only do that organically, but also with some of our investment that we're making through our ventures group. On the data side, Bill, that strikes me as an exciting, there's a lot of white space there. Maybe one area that I've been thinking about, and I'm sure others have too, is on the analytics side, where a lot of your customers are collecting their most valuable data in Salesforce. Today, when they want to make sense of it and query it, unfortunately, I'm oversimplifying, but they're extracting that data from Salesforce and moving it into some third-party query engine where they're drawing the value from it. That seems like lost wallet share and an opportunity for Salesforce to become a bigger player. Do you agree? I do agree. I think that there's an immense opportunity to help companies. We're generating data in a rate that we've never seen before. Helping companies understand the signal-to-noise ratio with that data is an important sort of place to play. It's why our Tableau business is so material for us. Tableau was such a pioneer in the world of explanation and the world of sort of visualization of what data actually means. And I think that the combination of Tableau plus our Data Cloud plus all these other diverse data sets we talked about gives this sort of opportunity for Tableau to have another rebirth moment, which I think is quite exciting. And again, it's not. I think a lot of this world of data and analytics today is still reserved for the few, the data analysts, the specialist sort of workforce in an organization. Our mission at Salesforce is to bring that to the many and how every sort of information worker in a company can get access to this kind of insight for the future. Yep. OK, interesting. And maybe I'll close. Sorry, I consumed all of our time, guys. But I'll ask one more. Bill, if you put your futurist hat on and think a little bit about five, 10 years out, which I know you get paid to do, besides spatial computing, which you mentioned a little bit earlier, are there any other really cool technology shifts that are super, super early now that you're watching that you would encourage this group maybe to just keep an eye on because they could be big in five or 10 years? Anything else that's kind of fun to flag for us? Yeah, I certainly see kind of building on sort of, we talked about cloud communication. We talked about data. We talked about business process. I think what will get really interesting is the verticalization layer of how all of this sort of communication data process encapsulates in a more verticalized model, and I think our early investments in our industry compute platform are proving some very interesting returns for the company, not only in helping companies have software that's purpose-fit for their industry, but also transforming the industry, and I think that, to me, is probably another area of exciting sort of category to monitor. A little bit like your Life Sciences Cloud. Exactly. Yeah, exactly. So I think those are the kinds of things that I would really be keenly tuned into, which is the verticalization and then the verticalization to then the localization of how that industry performs in a market or geo. That, I think, will be really exciting. So the world, I think, just gets smaller in some sense, gets more personal. And that's definitely the way I would be thinking of the shape of that curve. Bill, that was a fun conversation. Thank you. That's all the time we've got. Thanks, everybody, for joining. Thank you.