Salesforce, Inc. (CRM) Earnings Call Transcript & Summary

Good morning. Thank you for joining us today. I'm Mike Spencer. Today's session is going to be focused on providing an update on our customer momentum and customer success. As you heard yesterday on the earnings call, we're bringing humans, data and AI and apps together to build the Agentic enterprise. And today, you're going to hear an update on how this is translating into accelerating our customer momentum and fueling our relentless focus on customer success. I want to add that this is a webinar in the series of webinars we've been doing for the past several quarters with the intention of giving you more visibility and more insight into what's happening with the business. 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 in 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. And with that, let me introduce our leaders that have joined me today. Miguel Milano, who's going to start. Obviously, you all know Miguel Milano, our CRO; and then Srini, sitting next to my left, Srini Tallapragada, leads our engineering and customer success organization. I'm super excited. Prior to handing the floor to them, I'll encourage you to submit your questions online into the chat windows, and we'll make sure and try and get to as many questions as we can once Miguel and Srini are done with their opening remarks. And with that, Miguel?

Thank you, Mike. Thank you, everyone, for tuning in today. I wanted to use my time, Mike, if you're okay with it. First of all, trying to summarize at a high level what the quarter looked like Q3 that as far as I'm concerned, a quarter is a quarter check, amazing epic. But where we are [ miraculously ] focused is on the huge opportunity that is coming ahead. It's in front of us. It's the Agentic Enterprise. And I want to double-click with some slides, if you don't mind, to walk you through how we're thinking about it. So the quarter, Q3 was pretty epic. The reality is it was the best Q3 ever in the history of the company. It was actually the fastest growth that we did in bookings, but also in net new AOV since pretty much fiscal year '22. And this is -- obviously, this is the output. This is the result of what is really happening underneath. Very important, net new AOV, net new AOV, which is the purest measure of customer success, grew significantly more than the AOV. And as I -- as we shared in Investor Day, Robin and I, when net new AOV grows more than AOV, then AOV accelerates and ultimately, revenue accelerates. And this is what we all are obviously very focused on subscription and support revenue to reaccelerate. So we are very confident in H2, net new AOV is going to be several points ahead of the AOV growth. And we feel also very comfortable looking at the pipelines, looking at the capacity, looking at the momentum of the business that the net new AOV growth above -- several points above the AOV growth is going to continue. So one, you're seeing the cRPO numbers, which were amazing. Well, net new AOV grew more than that. And bookings for the quarter also grew more than that. So pretty exciting. But the most exciting thing is that the agentic enterprise opportunity is right here. Now it's very real. It's been rumbling for the last, I would say, months. We started feeling it in the demand in the bookings and the deals during Q2, but Q3 became very real. So let me explain to you what the agentic enterprise opportunity is because we've been -- for 26 years, we've done a pretty good job selling into the CRM SaaS market. We become the leaders in that market. We have 20-plus percent market share. We have more market share than our next 4 competitors together, next 5 competitors together. But in clouds like sales and service domains like sales and service, we actually have 40-plus percent market share. It's pretty unbelievable. But this is a multi-hundred million -- multi-hundred billion TAM market. All of a sudden, there is a new market coming at us, which is the agentic enterprise market. We define it as agentic enterprise, that is the digital labor market, the agentic market. And this market is probably an order -- potentially for orders of magnitude bigger than the SaaS market. And we've been invited to participate there. We happen to have the software infrastructure that is required to be successful in this new opportunity. And every company is knocking on our door saying we want to become an agentic enterprise. We've been experimenting for 12, 18 months, particularly the large enterprises that have the resources. They've tried to do it yourself in many different ways. And every single one of them, most of them get to a point that they get frustrated because for enterprise scale -- for enterprise AI to scale, they need the last mile. And the LLMs alone doesn't give the last mile. And we can, by the way, do this slide with 2 of us because we see it very clearly. What enterprise AI needs is you need the context. You need the data, what is happening in the company around your customer, but also the metadata, why that data is important. Then you need the AI that is able to do the probabilistic reasoning, not the deterministic reason. Of course, everybody has AI. We have it embedded in our platform. But then you also need the apps because you want the AI to do smart decisions based on the data and the metadata, but then you want the AI to trigger some execution, some workflows and you want them to be as deterministic as possible because you don't want big corporations to rely on AI to an agent, an LLM to decide how to return a good that was broken to a customer or how to correct a billing that was mistaken. You want that agent to follow the standard operating procedures that the customer has defined that have been qualified in the apps. The apps -- most of the customer-facing apps happen to be on Salesforce. That's a huge advantage. And then finally, humans. We all have come to a realization that AI is not replacing humans. AI is augmenting humans and humans with AI, humans with agents are going to deliver new levels of productivity, new level of customer satisfaction. You need them working together on the same workflows on the same platform. And this is our moat. We bring the last mile. And this is what companies are realizing. They've been experimenting. Many of them -- I was in a conversation yesterday. Somebody said, Miguel, I've been talking to some customers of Salesforce that are building the agentic layer outside Salesforce. I'm like, yes, there are many of them that try to do it, try to get the data because you can do it yourself anything. They get the data out of Salesforce, then they build the logic, then they decide that they want to take some actions, but they don't have the capability to execute the thousands of workflows that the customers have already built on Salesforce. So basically, by doing that -- and by the way, the humans are working here, and they built this agentic layer here, trying to build a cool interactive conversational UI, but a UI that is disconnected from the humans and it disconnected from the execution and is semi-disconnected with the data because the one -- the moment you move the data out of Salesforce into other data lakes, et cetera, it just becomes obsolete. You can do it, but it's expensive. So most of these customers have realized that we need -- they need the last mile. And they turn to us and they said, "let's do it together. Let's start the agentic transformation journey." So if you go to the next slide, what we've done over the last 6 months is train our teams to know what we call the agentic enterprise playbook to know how to capture this opportunity. And then we make them think in 5 steps. The first is present an industry point of view because ultimately, the industry point of view is the demand plan is miss a customer -- missed customer. We try to go as high as possible because the agentic enterprise is a Board decision. It's a C-suite decision. You want to transform, you want to be more conversational, you're going to be more intelligent company, you're going to be more proactive, you want to augment your employees, you want to increase revenues, you want to increase margin, you want to make your customers happier. Okay, you need to become an agentic enterprise. How do you do it? That's our industry point of view. I'm going to -- if you go to the next slide, this is what we've done. And this is like an [ iChart ] on purpose because we believe it is such an IP differentiation for us that we didn't want to openly share with everyone. But essentially, what we're doing is for every industry, we started with our top 10 industries. But by the end of January, we're going to have it for the top 24 industries. We've created a very detailed point of view on exactly for every industry, what are the key domains, what are the key workflows in those domains. And for those workflows, how those workflows are going to be reimagined with agents. And typically, for every key workflow, we have 4 or 5 agents. How those agents work together, we described it in our point of view, who is the human manager for those agents and how the agents interact with the human manager, how those agents impact specific business metrics of the company. And then how those agents are going to be deployed over time. We defined the agents. We created not just a value calculator, we created a database of thousands of agents that have all the definition, the role, the actions that they need to take, the data that they can access to, the guardrails, the things that they cannot do, the channels that -- where they surface, where we will surface the agents. This is really a very rich IP. When we put in front of the customers, they say, "Oh my God, we want to become an agent enterprise, let's start." We come in with Horizon 1, Horizon 2, Horizon 3, that's the demand plan. And at that point, they look at me, they look at our teams and say, "Okay, wait a minute, this could be expensive" because you guys have all these sort of metrics and consumption of data ingestion, conversations, seat base, how are we going to make this work? And this is one of the parts of our methodology to sell the agentic enterprise is we present to customers all sort of pricing options to meet them where they are. Some customers, we go -- I think there's a next slide. So I got to ask this question in the earnings call yesterday. And essentially, we didn't have any of that 9 months ago, a year ago. And we realized -- we listened to our customers, we realized that different customers are in different stage of their transformation journey, and they need different commercial frameworks to work with Salesforce. If you go to the right, that's my favorite one, okay? That's the agentic enterprise license agreement, AELA. You're going to hear a lot about AELA because AELA is essentially when customers are determined, they see the 150 different agents that they needed to deploy across their 20 processes, they just want to make sure that they have a flat fee that they have predictability on the spend that they don't have to worry about ingesting data or conversations. We are all aligned. It's a risk-sharing model between us and the customer where we say, pay me a flat fee for the next 3 years. Typically, it's a multimillion dollar incremental to what they're paying. And just let's start deploying. Many times, we have FDEs and professional services resources just to help them drive success very rapidly. And this is becoming one of the favorite commercial frameworks, Mike. And we sold -- we basically put it together at the end of Q3. We sold 16 AELA, pure AELA, then we had versions of AELAs like maybe dozens or hundreds, but 16 AELA of 16 customers that went all in and said, "Okay, next 3 years and in most cases, we double, triple what customers were already spending with us." But you can go to the left side, which is, okay, I still want predictability that I like the seat-based additions that you have. I just want to make sure that my employees that have -- that use a seat-based license for you, they have unlimited access to all the agentic power to be augmented. So all unlimited employee usage of Agentforce is included in the new additions, the Agentforce for sales, Agentforce for service or our, I would say, magical SKU, our top SKU, which is A1E, which is the Agentforce 1 Edition, which includes pretty much everything, Agentforce Data Cloud, Slack and many other things. In the middle, you have the pay-as-you-go. You are not sure, okay, let's start with pay-as-you-go. There are some customers that like in the hyperscaler world, they like that kind of relationship. The pre-commit and then basically, they can consume those credits over a period of time. And then we have the flex agreements, which essentially, if customers believe that agents are going to take the place of some humans and they may need less seats, which, by the way, we are not seeing in our customer base yet, but it gives them the reassurance that they can move seat-based licenses into credit -- flex credits for Agentforce and Data Cloud. So anyways, all these pricing mechanisms are really resonating very well. AELAs are resonating incredibly well. We have -- we are approaching now 100 AELAs. Every week, there is 10 or 20 more AELAs that are added into the pipeline. And just to let you know, we do AELAs in the low end of the market, mid-market and then the top of the market. But for my team to ask approval for an AELA, the size needs to be more than 0.5 million in the low end of the market, more than $1.5 million, I think, is in the mid of the market and more than 5 million of this is net new incremental. I don't know if I have more. Yes. The next slide is the NAV, which is -- it was also a key message of the call. So one of the key messages, again, the quarter was amazing, but let's focus on the future. One of the key message of the future was humongous agentic enterprise demand coming at us. This is like nothing we've seen before. Second message was we are uniquely positioned to win in this new massive TAM, which is multitrillion TAM. We are uniquely positioned because we have the last mile. And the last message of the call was something that we already shared with you in Investor Day is the lines have crossed. I told you -- when I was on stage in San Francisco, I told you the lines are crossing. I see the lines crossing. And when the net new AOV is above the AOV, AOV accelerates. And this is what you need to think about as we move into the future because we are confident that in fiscal year '27, net new AOV will continue to be a few points ahead of the AOV. We want to make those few points be many points. That's our goal. We are very confident. If we were confident, hopefully, we sounded confident in the investors call a month later, we are even more confident. Now the AOV growth acceleration will translate into revenue acceleration, subscription and support. Robin said it very clearly, 12 to 18 months from last month. I kind of joke with her and I saI'd, "Well, then it's going to be 11 to 17 months." But it's coming. And we are -- I mean, I think we have a reputation of being reasonably clear and conservative on how we guide. If we all are saying from the CFO to the CRO that the subscription support revenue acceleration is going to happen between 11 and 17 months, you could be confident of that, too. There is a lot of reasons why bookings are accelerating. The capacity, I have 15 -- I have -- today, I have like 13%, 14% more ramped capacity. In total, I have 23% more capacity. At the end of the year, I'm going to have 20% more capacity than a year before. I'm going to enter the year with 20% more capacity, of which more or less 15% growth in ramped capacity. Ramp capacity, they are productive capacity. The innovation that Srini and Steve have delivered, I mean, I'm probably the luckiest CRO in the world because I've never seen so much. It's not just Agentforce. It's even across our core clouds. We also launched ITSM. I think we're going to have a lot of good discussions in our next earnings around ITSM is booming already. We just launched it. We won dozens of customers. Life Science Cloud was a great example. It just quadrupled or I don't know the specific statistic, but it grew a lot during Q3. And I just want to mention that this is pretty big because we were doing a lot of business in life science and also medical devices, et cetera. And -- but as you know, for many years, we had this great partner that we basically partner and capturing that opportunity. We were capturing the outside commercial side, the opportunity. They were very focused on the commercial side and the clinical side. And then 1.5 years ago, they decided to compete head-to-head with us, and we decided to build solutions to compete with them, and the results have been incredible. I mean we announced yesterday Novartis. We had announced Pfizer of the top 20. Of the top 20, we've already won 5 or 6 and I say 5 or 6 because there's one that is in embargo, but I think there's going to be a press release later today, where we're going to announce another major pharma company going to us. So we are gaining market share from Veeva like there is no tomorrow. We've already -- in addition to the big 20, which there are still many in the air and they are reviewing, they all want the Salesforce platform. That's why they were happy with Veeva. They want our agentic capabilities. They want our data cloud to unify data across all the different domains in the business. And they are afraid to move with a small player that is going to build a new platform for them. So many of them want to stay and are staying there's more than 100 life science cloud -- life science customers that have selected Life Science Cloud and are moving off Veeva, and we are just getting started. So that's an example of innovation. Anyways, I'm excited about the future. I'm pretty sure that we're going to continue to show this slide because I like this slide that -- so that you see that net new AOV is above AOV. And that's my role. That's also my partner in crime role here at Srini, because he has all the resources for customer success, and he's obsessed in making sure that actually, we both are responsible for the net new AOV line. So that's what I wanted to share with you, and I'm looking forward to the Q&A, and then I'm going to hand over here to my...

Thank you, Miguel. I think as my role as -- first of all, thank you, everybody, for joining and appreciate the time commitment you have given us. My role is, again, as Head of Engineering and also customer success. So I think when the industry is changing so much, this is a period of transformation. And what we want to be very sure is when things are changing so much, we're very closely tied because the products are moving, the underlying tech stacks, the models, everything is changing. Key is to be very tight feedback loop with our products, with our customers and really feedback that loop. And even the pricing example you saw, we are listening very hard. We are going to fight very hard because I think a lot of things are changing, and we don't take anything for granted, and we are very hungry. And some of the examples, I just want to give you a couple of examples quickly on some customers and what they are seeing. And a lot of these customers have been -- have gone through this journey with us. And one example is Falabella. Falabella is a Latin American retailer. They are a big, very big company. And now what they did is like they were having this problem that a lot of their customers were calling their regular call center, but they wanted to redirect to WhatsApp. And one of the things they used Agentforce was Agentforce can work in any of the channels the customers are in. They turned on Agentforce. And really right now, they've increased -- they're doing about 216,000 monthly WhatsApp conversations, 60% questions auto resolved. And really interesting thing is adoption increase they're seeing. They -- as I said, the stated goal is to move more and more to WhatsApp and then they've got for 40% adoption increased just from August to October, and they're live in 3 countries in Colombia, Peru, Chile, and they want to go a lot more. And that's a great question of somebody who's really saying a retailer, a consumer agent, which is really answering questions.

I think on Falabella, I know them very well. Actually, I met the CEO a while ago. This is the sort of the Nordstrom for Latin America [indiscernible]. And I think what is pretty impressive is they started with a proof of concept, and they refilled the tank, as we said, they came back to us twice. And now it's a multimillion dollar relationship just in the agentic part of the business. Super successful, super...

And most of their things is in Spanish. They want to add more languages there. Now that's one sort of retail use case. Then if you look at Smarsh, Smarsh is a global leader in communication data and intelligence for regulated industries, very complicated use case with a lot of complex queries they get. They use Agentforce. This is where a lot of our advanced RAG and reasoning systems and how we do vector indexing. This is not like to make these answers very sophisticated answers. Again, this is other example in the telecommunication industry. They're seeing a 20% increase in customer success rates. They also tried to do it by themselves a couple of times. They tried it. They saw our system because there is a complexity that we can handle at an enterprise scale, and they're really seeing 20% increase, faster resolution, 30% increase in service productivity. There's like one another different industry, different case, class of problem statement. So if you go to the next slide, U.S. Telecom, French telecom company, major French telecommunications company. They've been a Salesforce customer for 10 years. They have 10,000 field service contractors. They use our service cloud, our field service, marketing, Data 360, MuleSoft. Existing Salesforce customer, then imagine they all have all the data already in Salesforce. They really say, how can I now change my thing? Their average call handling time was 2 minutes before. They really want to improve that. They used Agentforce service in the line of flow and then they reduced from 2 minutes to 12 seconds. And then their accuracy, again, when -- again, these are complicated use cases. And then they found a 95% use case. And then one of the things we also measure as a leading indicator is how many of their workflows that the customer has are agentified. So they are doing -- they're running about 3.5 monthly agentic workflows and they call their agent Iris. By the way, this is another thing we are seeing. All of our customers, especially who are public facing, they name their agent. Williams-Sonoma, we talked on the earnings call multiple times, they call their agent. Almost everybody gives it a personality, which is tied to the brand of the company. I personally think just like people have a website, going forward, they'll always have an agent, which represents as brand. It starts with one use case in customer support. Pretty soon, it goes to other use cases. And for all of this, again, to bring it back, the underlying platform is important, the context is important. Then you will see another in financial services industry. The financial services PenFed is a credit union, they got 2.8 million members, $25 billion in assets. They wanted to do their loan underwriting reviews, AI-driven loan underwriting reviews because that's, again, a complex regulated industry use case. And then they are finding that they can handle 10% faster handle time for reports, 75% of ITSM cases auto resolved and 30% projected cost in call centers. This is an example, again. Now if you really look at it between all the 4 examples, I just want to highlight a little bit of the last mile that Miguel has referred to is to do that, you need an underlying platform. You need the context and context has to come from different data layers, different ability to access data, structured and unstructured, really reason it, tied to the deterministic and nondeterministic workflows, really understands the jobs to be done, able to do not just getting live. Now these are all public facing. So they need their compliance, regulatory concern centers. They need a lot of security, prompt defense injection and all of that. And then day 2 operations, how do you monitor the agent? How do you eval, we call evals, but basically, it's testing and ensuring that the agents what they are, how do you observe them in production. So this is what is happening. So I just wanted to give you a sense of how we are very close to the customers and all of them -- some of these customers work with their own FDEs. Sometimes they implement it themselves. Sometimes they're implementing mostly with our partners. So they're taking the feedback. And based on these early customers, we almost added 120 features in the product. So we are learning very fast, iterating into the product, and that's the journey we are in. So with that, Mike, back to you.

Great. Thank you, both. So we are going to move to Q&A. We've got a number of questions coming in that we'll try and get through. [Operator Instructions] So with that, let me first start. Miguel, this is going to be for you. It's coming from Kirk Materne, and he wants to double-click a little bit more into AELAs and give a little bit more color on -- and we've been getting this question quite a bit on the impact of AELAs that they're going to have on net new AOV, like how it translates from agreement into net new AOV and AOV. And then more importantly or just as important, what you see coming on the horizon. So as you think about the second phase of the agreement, so they sign an initial 2- or 3-year deal, what does that look like? And then what's the shape of AELAs looking like in the pipeline as well because you guys have talked a lot about the pipeline.

Perfect. So yes, we're getting this question a lot. I'm very excited because I think people are understanding the power of the AELAs. AELAs pretty much cement the relationship with -- that we strike with our customers. So it's going to have a big impact on any attrition risk on any of the clouds that they had disappears. Typically, it's a significant increase, step change increase in how we are monetizing the customer because this is the very important thing. In the SaaS world -- in the CRM SaaS world, customers are using our product like a CRM. It's a SaaS CRM and the typical use cases. And we've been growing nicely, and we added some -- we added field service. We added Marketing Cloud, and we add 10%, 20% to the way we monetize the relationship with our customers as they add new clouds. The agentic enterprise world is totally different. It is a TAM that is multitrillion. And then when customers choose the Salesforce platform to transform all the processes to become the digital labor platform for all the transformation, the impact, the ability that gives us to monetize the relationship is a step change. It's a multiplier effect. It's not any more 10%, 20%, it's times 2, times 3, times 4. We -- in the Investor Day, we talked to you about x3, x4. We're already seeing it with a lot of customers. Obviously, AELA is already -- in most cases, is already double or tripling the business that these customers have been doing with us for 20 years. And all of a sudden, in 1 year because they want to become agentic enterprises, we put in front of them 10, 15 use cases, they go all in and then they double the spend with us or they triple the spend with us. I mean I presented Vivint as an example, Finnair as an example, at the Investors Day. Well, those customers started solid customers that were in the single-digit million relationship with us with multi-cloud, very happy. And then agentic enterprise came, they thought how can I transform should I go to this vendor, should I go to this vendor? And they all realized that they needed the last mile that Salesforce provided. I already explained the last mile, the differentiated. That's our moat. And they all bet on us. They started with the proof of concept. We came up -- I mean, I love the names of how our customers are given personalities to these agents that in the case of Finnair, Sisu, which means resilience is the agent that they put in production. And now many travelers that are stuck in airports and have issues, they call Sisu and Sisu resolve all the problems. And Sisu has been so successful. that the volumes of actions that Sisu is taking per week has multiplied by, I think the number is by 5 just in the last few weeks. So they went into an AELA with us. And Finnair, which was a stable company in just a year has more than doubled the relationship with us. And this is just for a few use cases. So they have many more to come. Giving is the same example, home security, smart home solution. Randy was with me on stage. He's really -- he is the Head of Engineering. He's really technical. He baked us off against all the possible agentic enterprise layer competitors. They chose us because he said, "I'm going to choose you because you have the data, you have the humans and you have the deterministic workflows that we want to execute on." He piloted. He launched 3 agents with 20 or 30 use cases for every agent, mega successful. Again, another 10x increase in volume of actions taken by the agents. And he said, "Okay, I'm going in." And then he did an AELA. And we more than doubled -- a very large relationship that we have with him over 10 years, we doubled it in just 12 months. And this is happening to every single one of our customers. So the second part of the question is, okay, now that you have the AELA, you have the visibility. It's a partnership, it's a shared risk, okay? We take some risk because I really want -- the worst case that can happen, the worst-case scenario for us in an AELA is that the customer is so successful that they deploy 100 agents that they consume a lot and we have to hit the LLM a lot that we had to ingest a lot of data and that the cost to serve of those agreements may be high. But guess what, the renewal comes, the customer has already developed 50, 20, 70 agents. It's consuming a lot. Those agents will need to consume even more in the future, and there will be another 50 or 100 agents. So we will recap the agreement, and we will make it, again, share risk, but we are very confident that the profitability of these agreements are going to be very high. So this is going to be a cornerstone of our strategy. For me, the message is this new TAM that we've been asked to participate in is a new sport. We are now playing into a different field, and it's significantly much larger. And every time that we get one of our customers, SaaS CRM customers into the agentic enterprise new sport, essentially, they double, triple or quadruple the business that they were doing with us, which is very exciting. By the way, they are very excited because they use our software in incredible new ways. I mean, the conversations that we have with these customers when they talk about their Vivun is Ava, then Sisu, then Oliv, it's very exciting. You're going to see a lot of household names. In 5 years, everybody is going to be talking about the names of these agents for every service, and there's significantly more to come. We have 18,000 customers. Nobody, please ask other companies how many customer-facing agents have you sold? Tell me stories. They will give you 2, 3, 10 stories. We have 18,000 stories and growing. By the end of the year, we're going to be close to 25,000 to 30,000 stories, okay? And we're going to start the year with those customers coming back and refilling the tank. I mean I was -- in Q1, if you remember, I was very excited because I found 3 customers. We had at that time, 3,000 plus, 3 customers, obviously, most of them have bought in Q4. In Q1, 3 of them knocked at our door and say, "Okay, I need more credits. I need to refill the tank." And I was very proud. Well, this quarter, in Q3, more than 50% of the bookings that we did with Agentforce, but also with Data Cloud came from customers refilling the tank. There were 362 customers refilling the tank, okay? So as the installed base of Agentforce customers grow, the number of customers coming to us refilling the tank is going to exponentially grow. And this is -- I mean, this is the [ panacea ]. This is the world of consumption where short -- I mean, the sales cycles become very short, where a productivities go up because they don't really have to do much. The digital labor is already working for them. And then the customer calls us, "Okay, I need more credits. I need more fuel for the tank." Anyways, very exciting.

Thank you, Miguel. So we've got a flood of questions coming in. This is great. I really like this next question from Terry Tillman, my friend from Atlanta. And this one is going to be for Srini. And personally I like it because we spent a lot of time as a leadership team talking about this, the importance of this technology as we continue to advance Agentforce. So Srini, could you give a little bit of color and introduce maybe for the folks on the call that haven't been as close to it, the voice innovation that we've been injecting and recently launched into Agentforce and the relevance of Agent Script as well.

Yes. Thank you, Terry. Great question. So as we've been working on voice for a year. We GA -- we did a lot of pilot customers. We GA at Dreamforce. Now we have a lot of customers who are trying it. I think the critical thing to understand is, one is voice is a new -- right from day 1, we realized when we build the architecture, the -- one of the most important things is the agents have to work in different channels. I gave you a WhatsApp channel, as an example, at Falabella like it could be WhatsApp, SMS, voice, web, embedded in other agents like ChatGPT. So it has to work on all channels. What customers do not want to do is redo it for every channel because you realize that it's very hard to get all your data right, put all the guardrails and do it and then imagine repricing it. So one of the big advantages of us on Agentforce is customers, when they try to open up in one channel, it's just switching a switch and enablement, and then it will work for voice too because the same logic, the same prompts you have given, same reasoning, same knowledge basis, everything the voice agent will be able to use because what drives the voice agent is the same reasoning engine that drives the chat agent, the WhatsApp agent and all. So that's why voice is there. Hopefully, we cannot share -- we haven't gotten permission from these customers to share their names, but we have a lot of customers, but hopefully, by next earnings call, we'll have customers who we can publicly reference as voice. Tied to that, you asked about Agent Script. One of the learnings that -- this is important, I'll keep repeating to that. We are working very closely with the customers. We are working. We are hand in glove. We are trying to learn what they're doing, RFDs are doing, our partners are doing. One of these things, and I think if you'll realize this word I use called what we realized is people are writing prompts. They say, do this, don't do this. They keep -- the prompts keeping longer and longer. And then the LLMs are great until some point and then they get confused. And so what we saw was our developers, our engineers or everybody, our implementers spending a lot of time in what I call a prompt doom loop where it's as though we forgot in engineering the regular programming. So I think it looks like a hammer -- everything looks like a nail because I have a hammer. So what we said is, hey, there are a lot of things. The LLMs are very good. Let's use them. There are a lot of things there, they're not good. So if you really want a prescribed flow where you always want to do some authentication, LLMs do it when you give the instruction, they don't guarantee it. In a B2C scenario, maybe 95% accuracy is okay. In an enterprise scenario, they want 100%, the standard operating procedure. And then what we found is, hey, there's a much easier way to do it. So we introduced this Agent Script, which says, "Hey, let's use the power and creativity of the LLMs where you want and the determinism of the regular script or regular programming when you don't want that." And in an easy way so that your time to implementation gets less, your testing cycles get reduced. So I think it's almost -- you have to think about it is Agent Script. Just like voice is new channels, Agent Script makes the entire agents much more resilient and for an enterprise use cases, but also make your time to test and go live much shorter. That's why it's very powerful.

Great. Thanks, Srini. Next question we have is coming from Mark Murphy and team, and we're going to double-click into the kind of the do-it-yourself dynamic that we've been seeing because I think it's a really important dimension to our ramp of Agentforce. And the question really revolves around customers who may start out on the do-it-yourself path and then coming back or boomeranging back to Salesforce in the end to adopt Agentforce. But I'll ask both Miguel and Srini to chime in on this one. But can you guys provide some color on what you're seeing in the customer behaviors? And then just as important with that, what's the time to value that you're seeing with your customers? And what are some of the things maybe for Srini, in particular, what are some of the things we're doing to help accelerate that time to value?

Yes. So the good news is -- actually, I don't like the word boomeranging because it's not that they leave Salesforce and then come back, which is sort of the boomerang. It's just that they believe that they can build the capability without Salesforce. And then later, they realize that they need to do with Salesforce. So first of all, the good news is, and this is very important. I keep referring you guys to the huge opportunity of the agentic enterprise. Every single company in the world, I mean, last quarter, I was in 12 countries, 3 continents. I met 400 customers. I had conversations with 400 customers, probably it was 100 at once and then every single one is experimenting with multiple technologies. In different domains of the company, they use different technologies. So obviously, the one domain where we really want to own and really grow with our customers is the customer domain. But even the customer domain, there are many customers. I mean we have 18,000 Agentforce customers out of 200,000 basically. So it means that the other 190,000 customers today, they're experimenting with something else, which you know what, is huge news because they know what they want, they are going to hit the wall because they're going to realize that they don't have the last mile. And I'll give you a very quick example, and then maybe you can also add to that. Huge bank in Europe. I'm not going to go to the country Europe, otherwise, because it's the biggest bank in that country. They -- great customer of Salesforce, 30,000 Financial Services Cloud licenses, 8,000 Service Cloud in the call centers. I mean like, obviously, many double-digit -- healthy double-digit million of AOV with us. And 1.5 years ago, they started experimenting and they -- somebody had the great idea that they could do this with Microsoft. And then they took all the CRM data from Salesforce. They went to Azure. They put it there, then they use Copilot to get the logic and they put Microsoft for like 30 engineers to work on that for 1 year. They didn't get anywhere because what they built was disconnected from what their 43,000 employees of this bank were doing every day, triggering workflows. This bank, for instance, they have more than 1,000 different flows that have been codified in the Salesforce apps, which is our automation that the humans are triggering every day to the tons of millions of automations triggered by these humans every month. And so they built something that was disconnected that was super heavy, that was very cost to maintain. And then at that point, obviously, we have a great relationship. And we say, "Guys, we can do this." It's embedded with the humans are with the best technology, with Agentforce, with Data Cloud, bringing all the information from the different units together and they say, "Okay, let's do it." We put 3, not even FDEs and solution engineers, and they built the same but embedded with the humans, with the workflows in 1 month. And then a month later, we signed a double-digit multimillion agreement with them to add to what they were doing, and that was just the beginning. Now we are talking to them, again, to double that because now they want basically all these capabilities to every single employee in the bank. And by the way, they are -- they love the other provider for many other things that it doesn't make sense to do it yourself when you are in the customer domain. Now if you want a solution for some supply chain or whatever, well, we were not a player there. By the way, now we are because we bought Regrello. But my point is there is a lot of experimentation. This is demand that is going to come to us. It's the best thing that can happen. 100% of the customers are going to become an agentic enterprise. Who are they going to choose? Listen, I think most of our customers are going to choose us. And what I said earlier, this is very important. When they choose us, it's not that our business with them is going to grow 10% or 20%. Our business with them is going to triple or quadruple. So do the math, of course, it takes time. I first -- we are first educating our thousands of new AEs that we are adding to our go-to-market. Then they have to get the FaceTime with them -- with the customers, be in front of them, explaining the concept of the last mile. And then the customer needs to make a decision and then we need to pilot and then we need to find the success and then they go big with an AELA. And it takes time, but we are seeing this accelerating like never before. That's a testimony of that is the pipeline. We never had this size of pipeline, very healthy double-digit growth in pipeline in open pipeline in next fiscal year. So things are moving in the right direction.

So just -- I think Miguel answered most of the questions. Just a little bit more specific context is, like I said, some of our most advanced customers, most forward-leaning CIOs, I would say, who are ahead of the curve, they're always innovators, they started 2 years. So they went through all the experimentation. So they are more clear on what works, what doesn't work, what is the cost. And so they realize that there are things which is not worked for them to solve. That's why you need a platform. This is the same -- why do you buy a platform versus why do you want to build yourself? I think they also got fed, including me as I'm also an engineering -- I'm a practitioner of whatever I do. I'm also using white coding my engineering team. I really know what works, what doesn't work, where is the limit. So I think they learned it. So they are much more clear that what it works. At the same time, by the way, like Miguel said, they have multiple agent initiatives. One of the other things which really helped us is across the platform, we standardized on open standards, right from the data lake layer with iceberg, right from MCP, different models and now with Agent Graph, MuleSoft Agent Graph, like where you do -- we support A2A. So it allows them to say that, "Hey, I can build specific domain agents." I can -- even including our observability with OpenTelemetry standards and stuff like that. So because the platform is open at every layer, now these customers are saying that, "Hey, for these domains, it's not worth me trying to build. I'm going to use it." But one of the things we want you, Salesforce is if I want to call your agent, for some reason, let's say, I built it on AWS or Azure or my -- custom Google or something, can I talk to your agent? And that's what MuleSoft Graph is. And I think one of the most resonating things which I found from these enterprise CIOs who are a little advanced is they already see the future of this multi-agent orchestration. And one of the most important things they glued on right now is MuleSoft Agent Graph because it allows them the federation. They can -- they don't need to do things which they don't need to do. They get out-of-the-box templates and agents, like we said, for each of the verticals with specific domains. So they don't do the heavy lifting when they don't need to do. They will definitely use their teams bandwidth on specific use cases, which are unique for them. And so we enable this multi-agent world with open standards, and that's what we are seeing. And I think some of the other people who are a little late to the game, they will come 6, 7 months. They'll learn that, "Hey, it's not the day 1 problem." When you go show a good demo, it's -- once it's in production and day 1 and day 2 because these things degrade if you don't do it. There's a lot of work just like a human, you need to performance manage them. It takes a lot. This is what we are learning with help.salesforce.com, too. They realize the overhead is not worth it. I better depend on a platform. So I feel that's what we are seeing, and I can see that going more and more.

Perfect. So we've got a number of questions coming in. I'm actually going to do kind of a two for here, and I'm going to jump down a little bit and then come back and combine a couple of questions. But -- so this next one is from Tyler Radke of Citi. And I'm going to take the first half of this question where he asked about the net new AOV growth statement of net new AOV being higher than AOV growth in H2 and what the comparison is on Q4 versus Q3. We obviously haven't disclosed Q4 versus Q3. We've been vocal about Q3 being strong. But the forecast and how we come up with it, obviously, is based on pipeline and our projections around expected bookings in Q4. So that also feeds our CRPO guidance for the quarter. So that's kind of the way I would frame that. The second half of Tyler's question, I'm actually going to combine with Michael Turrin's question from Wells. And it really revolves around the promise or the visibility we gave everyone at Investor Day around the uplift that we're going to see over the long-term horizon ARR AOV synonymous and what adoption looks like within customers for Agentforce and we go from first use case, second use case, third use case. Miguel laid out, obviously, a slide earlier that we consider to be very important to how we land this with customers on introducing a number of agent use cases. But maybe we'll start with Srini and Miguel, I'm sure will chime in. But maybe, Srini, can you give some insight into how you approach the deployment for deployed engineers, obviously, has become an important leg of our strategy in trying to drive that. And then Miguel, maybe if you can expand on that, talking about then what the conversation becomes with the customer and expanding to the second, third, fourth use cases.

So I think let me -- also there's a lot of hype around the term forward deployment engineer, like what it is, what it is not, like -- so let me just demystify some of this. When a new product comes, we always used to do that. We'll have a small team, which will work initially with the -- we would call them pilot and then have the engineering teams work closely to mature the product. Basically, in the world of agentic enterprise, with some of the underlying swirl is happening even in the underlying technologies, we wanted a little bit bigger investment to iterate. So one key role of our forward deployment engineers is to mature the product, okay? So that like it's easy for customers to implement, get the feedback, iterate it. So which I would call them as product FDEs whose job -- main job is to give the feedback to the loop, give feedback to the product engineers, what's happening in the ground, really iterate fast to mature the product so that we can scale it because we have 200,000 customers across so many countries. And then part of doing that also is work with our partners, SIs, correct, to really ensure that this is -- this agency transformation is like it's a slightly different model. So I think we are working with standardizing the playbooks and all for our customers. So why -- so as you see, now what was interesting for me is of the thousands of customers we have live, initially, maybe 8 months back, almost every implementation we were involved. Now a lot of times, I'm finding that this customer is live and very big usage, and we are not even involved, which shows which is where I want to be because I want to mature the product. So the cost of implementation can be -- first of all, a, the customers could implement it themselves or any of our partners can implement it, correct, while we are improving the product. So I think they're structured. And at some point, right now, for the initial customers, our aim is adoption and maturing the product and tracking it. Some of this we are bundling. At some point, if you are in AELA and all, we are going to be part of the package. And maybe at some point, we will also charge for the FDE motion. We are not there yet, but I think that's what we are doing. But we are moving -- my main goal is to really mature the product where we can really iterate it and then we get all our -- lean on our partner ecosystem. And in fact, we are working very closely with a lot of our big partners like Accenture, Deloitte, KPMG and others and across the world, specialist partners to really scale this so that they can take it because to really deliver this agentic transformation, the TAM, we can't do it alone. We really need our entire partner community, and we are having regular meetings that are part of our training. In fact, we're running FDE boot camps where the partners are coming with us. So I think this is going to be a whole of not just Salesforce but our entire partner ecosystem. Another thing I just forgot to tell you is because of these open standards, when we say partners is SI partners, but a lot of these agents, I think if you remember what Miguel has showed on the second slide of all the different agents, a lot of our ISV partners, our independent software vendors partners, they will build just like App exchange, there'll be an agent exchange. They'll be building a lot of the last mile agents on the platform, and we are working with a lot of them, and they will build those actions and agents. We already announced at Dreamforce hundreds of actions through our ISVs. And I think as the platform is maturing, we're taking the feedback and the whole cycle will repeat.

In fact, in the point of view slides that you saw a mini version, just [ copy pasted it ] in the bigger slide that essentially, when we go to every industry and we show the 150 agents that we need to deploy, we actually qualify every agent, and we said these 50 are embedded already in our platform out of the box. The 25 come from our ISV ecosystem that we've already certified them, so they work on our platform. And then these other 50, you need to build them yourself, but this is how you build them. This is the data that you need, et cetera. So partners are fundamental to really accelerate this revolution. The demand is going to be huge. Every -- one of our big partners from Deloitte to Accenture to all of them, they are getting ready with massive hubs with experts. They're getting all certified on the Salesforce platform, the Agentforce, the Data 360 and our apps. And I think one -- to finalize with the question is how fast people go to double to triple to quadruple, it actually depends. It depends on where the customer is in their mindset, their conviction on how much have they experimented. What I can tell you already is that we've seen dozens of customers, probably hundreds of customers already that they've doubled. But they've doubled a successful relationship that they had with us for 10 years with multi-cloud and in just 1 year, for an initial set of use cases, they doubled their spend with us which is a great indicator that probably they're going to go 3x, 4x, 5x. I mean there is one -- a group of Vacasa, for instance, is an example of a financial services company in Central America. They just went 5x in the first shot because they saw it. They want to go fast, they want to be differentiated. And they ---basically their relationship with us has multiplied by 5. So this is like a step function for our business, also for our partner business. And my goal, listen, is to move very fast to get as many customers to adopt our platform to become agentic enterprises and to do it with full commitment with an AELA. They can do pay as you go, and then we're going to see the x2 and x3 and x4, it's going to take longer. And of course, the very big accounts. We have a lot of accounts that pay us more than $100 million. We have a lot of accounts that pay us more than $50 million. Those will take a bit longer to multiply by 4 or by 5. But some of the wins, when I look at the top 10 wins, and this is a cool statistic that I would like to, if you don't mind, share with the team here. So if you look at the top 10 transactions that we did in Q3, just order of magnitude, those included more or less $105 million of net new ACV. And if you look at the TCV, so what is committed in the contract, approximately more than $1 billion, okay? The 10 transactions, basically in our RPO, it shows basically more than $1 billion. Well, in those transactions, in the ACV, 7 of the 10 transactions were driven by the agentic enterprise transformation with Data Cloud and Agentforce, 7 of the 10. There were others that were single cloud that they wanted to expand, but 7 of them, which is great because it's 7 of the top 10. And when you look at the -- how much did they bought of Agentforce and Data Cloud, I think they bought quite a bit, 30%, 35% of the ACV was Agentforce and Data Cloud. The rest were our core clouds. And in the total bookings, it was less than 15%. 85% were our core clouds. What this means is, number one is Agentforce and Data Cloud are prevalent in our biggest deals. Customers are accelerating their multi-cloud transformation because of Agentforce and Data Cloud. So Data Cloud and Agentforce make all our clouds better. That's why Mark continues to say, which is true. Sales Cloud is no longer Sales Cloud. Service Cloud is no longer Service Cloud. Marketing Cloud, it's Agentforce's Cloud, Agentforce Service Cloud, Agentforce Commerce Cloud, Agentforce Marketing Cloud.

Yes. Well, thank you, Miguel. Thank you, Srini. And with that, we're at time. I really, really appreciate the participation today and the interaction. The questions were great. If you haven't gotten enough Miguel, we'll have Miguel on stage at the Barclays Conference next week with Raimo. And then I do have an ask for everyone on the call, if there are -- what other topics you all want to hear about, we have a quarterly session of this going on. And so we're always looking for new topics to bring you guys more insights into what's happening in the business.

I know we are up on time, but there was a question about Commerce Cloud. And I think I want to address it because I don't want to leave it hanging.

Please, yes.

So we totally transformed our Commerce Cloud offering. The question was about what -- how are you doing to differentiate it? You seem to have lost market share, deceleration. So we've actually spent the time innovating and building a stronger Commerce Cloud. And the basis of our stronger Commerce Cloud is we made it headless. Most of the other options are also headless. We've added a very powerful order management platform around it, which is that business is growing very fast. Very important. We made it totally agentic. So it's not -- Commerce Cloud is not Commerce Cloud anymore. It's Agentforce Commerce Cloud. So now you're going to be -- you're going to have a shopping assistant agent on everything that you do. We've added a point-of-sale solution, very modern point-of-sale solution that is growing now a lot from a small base, but it's growing significantly. And then finally, we have connected it much tighter with the other clouds. So listen, if you want to just buy a stand-alone Commerce Cloud for our website, probably we're not the right option for you. If you want to build another touch point in your customer engagement, connected with marketing, connected with service, connected with Agentforce, we are the obvious choice. And that's the big deals that we're winning on Commerce Cloud. By the way, Commerce Cloud was one of the fastest-growing clouds in Q3, also for net new AOV, but of course, of a lower base than the other clouds. But I'm very excited about the future of Commerce Cloud.

Great. Well, thank you, everyone, for joining. Appreciate it. Have a great one.