EP 149: Marc Benioff (CEO, Salesforce) Predicts Half of Conversations Will be With AI Agents Next Year

0:00There were more than 100 million leads that we have not called back at Salesforce in the last 26 years because we have not had enough people. So we just couldn't call them back. But we have now an agentic sales just for our own company, more than 10 ,000 leads a week right now, having conversations, turning them into pipeline. Our pipelines have never been more full. And a lot of it has to do with that we're feeding them with this agentic capability. And so our salespeople are also now part and parcel in partnership with these agents as well.

0:32Mark, thank you for doing this. Oh, yeah. It's great to be with you, Logan. So I think it's been eight months or so since we had the last conversation. I think we're seeing a lot of, at least I'm seeing a lot of end of software. This is the end of application software as we know it going on, I guess. What's your perspective over the last eight months and some of the rhetoric about this, that we're no longer going to see the commercial opportunity for software that we've seen in the past? Well, I think there's two things. You know, you look at the last eight months. I mean, it's been eight of the most exciting months, I think, of my career.

1:08When you look at what's happened with applications and with software, it's pretty awesome. If you remember, our conversation was all about AgentForce, and we were getting ready to deploy AgentForce on our support layer. And now we have. And I just tell you, like, we're customer zero for our new Agentec service and support product at Salesforce. So we have now done about a million and a half conversations with customers. And at the same time, that's the agentic layer speaking to the customers. A million and a half conversations also happened through our support agents during that same period. And the CSAT scores were about the same, which was stunning.

1:49And also, I was able to rebalance my headcount on my support. I've reduced it from 9 ,000 heads to about 5 ,000 because I need less heads. But there's also an omni-channel supervisor now that's kind of helping those agents and those humans work together. And this is the most exciting thing that's happened, you know, in the last nine months for Salesforce. And we also are deploying thousands of customers with the same vision. I want to dive into some of those things that you brought up in a bit. But I guess I'm curious, the agent force, and you mentioned the cost reduction or the ability to kind of manage costs within customer support.

2:27And I think engineering as well, you guys have been able to manage what the headcount and scalability of that looks like. Do you think the biggest benefit of AI is going to be cost optimization for organizations? Or do you think that ultimately we're going to have revenue uplift as well? No, definitely not. I think that, you know, we are all on our path and I feel like I have to pioneer this as my, as I said, my own customer zero. And I just gave you the support example, but I have to tell you a good story, which is that, and something we did not talk about nine months ago, which is that there were more than a hundred million leads that we have not called back at Salesforce in the last 26 years because we have not had enough people.

3:08So we just couldn't call them back. But we have now an agentic sales that is calling back every person that contacts us in our company. And we're doing about, just for our own company, more than 10 ,000 leads a week right now, having conversations, turning them into pipeline. pipeline, our pipelines have never been more full. And a lot of it has to do with that we're feeding them with this agentic capability. And so our salespeople are also now part and parcel in partnership with these agents as well. So just in the way that the support people, nine months ago, I said, we were about to do that.

3:45We now have also done that in sales. And I have other examples I can give you because I am on a mission to make Salesforce an agentic enterprise. And I think that this is what can really transform every company. As you can see, I've rebalanced my support headcount, as I said, so I can now put those heads into sales. So I've increased my distribution capacity. And now I'm also making sure that I have much more efficiency and productivity in my lead generation and in the ability to actually work with customers that are contacting me. Even in my website, I have deployed an agent on the front of it.

4:22So if you go to our website, you'll see we've taken all of our data from our website, put it into our data cloud, put an agent on the front of our website. Now you can talk to our website right through our agent as well. And so that just makes the marketing so much more efficient. This is really the beginning of every part of the company having this kind of agentic augmentation. It's a force multiplier. It's a synergistic effect between me and the agents. And we're proving it out first at Salesforce. But we also have thousands of companies that are also deploying this. You talk to a lot of investors and you've seen a bunch of technological trends, I guess.

5:05Do you think public market investors or investors at large are sort of waiting to see this come to fruition, either in the growth rate acceleration or profitability before they really lean in on the themes? Or what's been the sentiment as you sort of talk to people in the public markets on the investment side? Well, I think you definitely see a lot of confusion with customers or with investors on what does it mean that we're moving into an agentic era? Does it mean that software and applications go away completely? I mean, that isn't true for me. I don't know if it is true for you, but I'm using applications on my phone.

5:42I will say, I mean, this is a thought exercise I was having the other day, and I'm curious your thoughts on it is like, you know, I've been long, my career's been sort of predicated on investing in software, broadly speaking, right? But I'm definitely going to websites less, like in using OpenAI or Cloud as the portal of accessibility in some ways. And so I was thinking, like, will I ultimately use some similar interface? And not to say that that's going to capture value that you wouldn't otherwise capture, but I do wonder about the interface element and the usability changing. Well, I think at Salesforce, what we have really seen is that in the type of applications that we're deploying and the automation that we provide to enterprise professionals.

6:24So I kind of gave you the example of, you know, our support center where we have an omni-channel supervisor that's managing humans and agents together. There's no better interface that I've seen, or I would use it, on top of our data foundation. Our data foundation is our data cloud, MuleSoft, Tableau, and Informatica. That's the core of our company. And we have a huge amount of data ourself in there, but we also manage 230 petabytes of that information for our customers all over the world. The second layer is the application foundation. And some of these areas, like support or sales, are highly specialized applications and require, you know, incredible, you know, powerful user interfaces.

7:06And by the way, it's not that I don't have like a Slack user interface for service and support or a Slack user interface for sales. But then, you know, I still have to branch out because all of a sudden I've got a lot more data I need to consume. And humans are not going away. We are working in partnership with these agents. And that's how I look at it. And then when it comes to Slack, like, for example, since I talked to you nine months ago, I now have, you know, a couple dozen agents running inside Slack on my behalf, doing renewals, you know, doing support, but also like doing my wellness benefits, doing, you know, all kinds of interactions with my employees.

7:46They operate inside the Slack application. You can see in the Slack marketplace all kinds of new agentic apps and ISVs that are running inside there. That's amazing. So that is probably the most exciting, most thriving ecosystem in tech for the enterprise right now is inside Slack. So I think we're really have a good, solid, you know, look at that every single day. Do you think this is maybe tied into my job in some ways, but do you think I don't know what percentage you guys have of the CRM market? I used Claude before this and it said 26 % was the number I got and ServiceNow was 44 % of ITSM or something.

8:29One question we've been wrestling with is like, do you think we're going to see just much more federation and like this generation of companies that are being founded right now might pick off a percentage point or two here or there? And the market overall will grow. And so it's, you know, a rising tide for everyone. But it's going to be really hard for net new businesses to have that level of penetration that you guys and ServiceNow and Workday and whoever else have been able to achieve. Well, I hope you'll come to Dreamforce and you'll see, actually, we've never been in the ITSM market before.

9:02So we're entering and we're building on Slack. You know, so it's a Slack-first app. It's a Gentic also. So it's a Gentic. It's Slack-first. And I mean, ServiceNow is a great company. They, you know, I think they automate like 9000 companies. I don't know what the number is that they have, but Slack is on a million companies. And those million companies, they're, you know, small and medium and large and extra large because all, you know, Salesforce sells to the five major segments of business at the all of the IT world. And Slack is one of those products that we do that with. And that idea that we're entering the market for ITSM, but it's, you know, a whole different kind of architecture, but it's the architecture that we're talking about, which is we have to have humans and agents working together.

9:47And that when we're, you know, managing, you know, our assets in our company or, you know, really kind of providing that kind of, you know, database to really understand the change configuration, you know, agentic capability is very powerful. And then if you can provide it right on your phone in Slack, that's really a next generation UI. And we're doing that with all of our products. So all of our products, like our sales cloud, which we've talked about for decades, it's becoming agentic sales. Our service cloud is agentic service. Even our marketing cloud, we do 11 trillion emails every single year.

10:25Those are all one-way conversations today. You're going to see a dream force. Those are all going to become two-way conversations. And it's because you're going to have an agentic on the end of that, having a conversation with those customers. And I'll tell you, like, even, you know, in my home, I have an Airstream trailer outside and it's hooked onto the power grid, you know, with this device from a company called Eaton. And they've been a customer of ours for years. They've done, you know, their sales and service on us. But we have this field service product that we also now have made agentic.

10:57And so when the field service agent from Eaton shows up, not only are they like on their phone, you can see the app, by the way, it's on the app store. Not only are they looking at all my information, oh, it's Mark Benioff, here's his home, you know, here's his Airstream, we got this working, or I don't understand why this is not working. He's not by himself, he's with the agent, he can interact with the agent and the user interface exactly as you're kind of describing that kind of, you know, large language model interface and trying to figure out what's going to work, what's not going to work.

11:25And then all of that is going to get stored into the system of record so that when the next agent comes out to fix my power system again, it's all right there in front of them because that's going to be the nature of field service. Even like when we work with PG &E on their wildfire prevention, they're out there with our field service app where they're looking at every power pole and every branch and every piece, but they're not alone, the agentic layer is with them. It's humans and agents working together. But the system of record, the agentic layer, the LOM, all have to be integrated together.

12:00You know, it's that, we talked about this actually nine months ago, I remember, the data layer, the data foundation, the application layer, and the agentic layer are the three core layers of how applications are working going forward. And that's, well, at least that's how I use them every single day. And maybe you have a different way. Even when I'm using, for example, like last night, I was shopping for ties. So I was, you know, on an LLM and I'm looking for the right tie of this kind of thing. It wasn't one of the two that you mentioned. It was a different one. And then it brings me in there.

12:32It's kind of a super search. You know, it's like helping me search to find the website. And then I'm now I'm on the website, but I still had to go to the website to complete the transaction, to identify, you know, who I was, complete the commerce. Will that maybe get more refined? It probably will. But when it comes to the fundamental trust, security and information sharing in an enterprise, that is a different ballgame right there. What is managing agents and managing workers entail? You guys have gone through this transition on the support side a little bit and increasingly maybe on the R &D and development side and now on the sales side as well.

13:10I guess, how do you think about if you're a leader and trying to figure out how to deal with both the change management process associated with going through that, but then also, like, how do you actually organize and design the business such that KPIs and accountability is actually roiling through the organization? Well, we talked about that too, nine months ago, which was, I really felt I was going to be the last CEO who was only managing meetings. Remember that? Yes, I do. And I still haven't come up with a better term of like managing agent workers or something. For some reason that feels a little dystopian to me, but it's clearly headed down that path in some way, shape, or form.

13:49Now, I don't think it's dystopian at all. I think that this is reality, at least for me, but I'm managing agents and humans. And that means my agents, when they show up to a customer, like I told you, I have that salesperson now that's calling back all my leads. And that sales agent, there has to be a certain kind of set of guardrails, a certain personality, a certain kind of tone, the ability to know when it needs to escalate immediately to a human being, because these large language models can do a lot of things, but they cannot do everything. And you know that. And that we're kind of hitting in some ways, you know, a high evolution of the large language model, as we can kind of see as we got into GPT-5, looks a lot like GPT-4.

14:32It's more evolutionary than revolutionary. So we're seeing a maturation of that type of technology right now, that deeply integrated into the application layer, that's where I'm getting the most value. And that idea that those agents then when they're directly talking to the customer or directly talking to the employee, they need to be managed through those guardrails. They need to be managed through that structure. And that's incredibly important for us as the tech vendor to be able to provide that to the customer. When you say 30 to 50 percent of work is being done by agents today, I guess, how do you think about the unit of work being done?

15:12How do you assess that at an organizational level to try to figure out where the balance is of who's doing what? Well, I think we kind of lay it out like in that support example, which is really, I think, the best example, right? that if we were having this conversation a year ago and you were calling Salesforce, there would be 9 ,000 people that you would be interacting with globally on our service cloud. And they would be managing, creating, reading, updating, deleting data. We called CRUD database with an app, a very powerful interface with Lightning. In some cases, if there's a case, maybe it gets swarmed onto Slack, but this is what they're doing.

15:51Maybe they're checking the analytics in Tableau. Maybe they're over here. you know, doing some other kind of sentiment analysis. Now it's a different system. Now, all of a sudden, here we are a year later, and the million and a half conversations that are happening going forward over the next year, and a million and a half human conversations that are going forward over the next year have now bifurcated. 50 % are with Asians, 50 % are with humans. And there needs to be also technology and people who are managing, how are these working together? And because all of a sudden, like the agent will scale out and say, whoa, I can't handle this.

16:30I need to, you got to go handle, have the human work on that. It's not any different than have you been like in your Tesla and all of a sudden it's self-driving and goes, oh, I don't know, I actually know what's happening. You take over. And that's kind of the same thing, but on a very complex, you know, you're working with a very complex customer, you know, in a support environment. That is really the magic going forward. You know, how do we get these two? And so that idea of the 50-50, I think that's definitely where we're going in sales, in service, in marketing, in field service, in the fundamental collaboration of our employees.

17:08All of these pieces are going to have that kind of balance. I don't know if you've reported since May on numbers related to this. What I had down was 8 ,000 deals and generating over 100 million in ARR for Agent Force. I don't know if there's newer numbers out there, if that's the latest. Well, you'll hear those when we do earnings next. Okay, great. Okay. So earnings, we'll wait in bated breath for that. I assume they're higher than they were in the past. What was counterintuitive or as this has ramped up, what's been unexpected in the response of actually seeing this out in the wild and in production?

17:44Well, you can see that our AI and data product line is the fastest growing product line we've ever had. It's, you know, it's over a billion dollars now. And that has been a huge shock that we've been able to scale it as fast. I mean, we've seen these great startup companies like Databricks and Snowflake. You know, we've been investors in both of those companies. We help bring Snowflake public or investors in Databricks as well. And, you know, and then, you know, we look at other companies that are doing data foundations or data foundries and others. is like, wow, this is a big category. And these companies are all like$3 to$4 billion in revenue.

18:19And now all of a sudden, whoa, we're a billion dollars and we're fast tracking to$2 billion in revenue on that product. So that's very exciting. I think that that idea that the data foundation product line, the data cloud, you could even say agent force at some level, though I wouldn't put it at that stratification. MuleSoft, which also provides the integration capabilities, as well as Informatica, which, you know, as we've announced we're going to acquire as part of our data foundation. And getting all of that working to really get our customers to harmonize all their data, because if you don't have that data together, you can't have the accuracy with the AI.

18:59This is really a huge category for customers. And I think that this category will radically expand or it has been expanding. It will continue to expand for us. Have you guys not closed Informatica yet? These big regulators, man, it's crazy how long these deals take to close. I guess you've never acquired a company before. Not at the scale, not at whatever your guys' valuation is. We've done quite a few acquisitions, I think maybe more than 60. So I'll just tell you that, you know, we just let everyone work at their own pace. Yes, yes. That's very diplomatic of you. I have a lot of respect for everyone.

19:32There are a lot of hard workers out there. Yes, yes, exactly. That's very diplomatic. They talk about a place we could use some agents, I guess. I guess one of the things we touched on last time that I'm very curious about is like the per seat pricing, which you guys were a pioneer in, and then the shift, the more agentic pricing. I guess any updated observations, thoughts, perspectives on like what's going to be maybe cannibalistic as you don't need as many seats to do it, but might increase overall dollars that a customer spends? How do you think about that now with a little bit more data? Yeah, I've been on the road for two months with customers.

20:08And so I'm just back now in my office. So, you know, my view of that really continues to evolve, which is, of course, with per seat pricing. We've always had consumption pricing also, which is like we saw in our commerce cloud or even on email or our data cloud is consumption pricing. Per seat pricing is like the sales cloud, the service cloud, Slack. You can see per seat pricing like on ChatGPT that you mentioned or, you know, these kinds of products. So per-seed pricing, consumption pricing, and then conversational pricing, or this kind of idea that maybe you have a certain amount of conversations or a certain kind of flex credits.

20:45And one more category, which is customers want kind of the ability to buy all of it as one big package. Way more aggressively than I realized, they want kind of a complete agentic enterprise license agreement. Every company is really on a path to become an agentic enterprise. I've articulated us as customer zero. Like I gave you the story now of service and sales or what I'm doing in Slack where I have a dozen agents and I'm renewing customers. You know, nine months ago, I had no agents deployed in my company when we talked before. You know, now I have lots of agents. And I think every company is going to, when they see what we've done and when you get to Dreamforce and see my keynote, I'm going to have 12 Fortune 100 companies there, you know, all showing you how to build an agentic enterprise.

21:32different shapes and sizes, different ways to do it, ranging from Pfizer to FedEx to OpenAI Anthropic. They're all going to be a dream force, really showing customers how to do that. That's October 14th through 16th in San Francisco. And I think that will be a very critical thing that I don't just articulate, hey, here's our great new product. Here's, like I said, the new ITSM product or this new product or that new product. But here is customers being successful, becoming agentic enterprises. Last year, we were just talking about agent force. We've talked about that. Here's agent force. You can deploy this.

22:09Go. Now we're saying, no, actually, this is a huge enterprise transformation for you. And in every category, we're going to transform that function. And we've had to rebuild every single one of our products to not only be the application layer, where you're creating, reading, updating, deleting the data, but also working hand in hand with the agents as well. And that is, I think, really exciting. I think for a lot of customers, they haven't been able to see it yet. And they want to see what other customers are doing and how to make it work and how to deploy it. And that, I think, will be probably my greatest pleasure at Dreamforce is to see customers show other customers how they're doing.

22:49You mentioned OpenAI and Anthropic there, and we touched on a little bit of the scaling elements of the model today. I'm sure you get to see a lot of stuff in preview or get to have conversations that a lot of people aren't privy to. Do you think we're increasingly moving to like an absorption and digestion phase where we just try to take advantage of all the value that we've already presented or those model companies have already presented for enterprises to take advantage of? Or do you think we're still on a fairly steep slope and it's a little bit of ways from the harvesting stage? It depends what category of the market we're talking about.

23:29You know, the market is really in five segments. You have the small and medium business, you know, call it a couple hundred employees. You have the medium businesses, call it a couple hundred to a few thousand employees. You know, you have the large businesses, call them like four or 5 ,000 employees. then you, of course, have the very large Fortune-wide editor businesses and you have the government. And these five segments of the market each need a slightly different version of the product. They need something with a slightly UI. They consume it a little differently. And the speed of their deployment is different because their architectures of deployment are different.

24:04And also, AI is going to dramatically change small businesses. There's going to be a radical explosion of the number of small businesses because AI is going to make entrepreneurs so much more capable and probably a media business too. And of course, in large business like mine, it's happening, but not everyone is probably as committed to deploying the technology rapidly and at the bleeding edge as I am. So those five segments will all adopt the technology slightly differently. And that's my observation of technology over 40 years. and so where do you think we are on that do you think we're at the absorption stage for for the mainstream enterprises or we're at the beginning of the beginning a lot of for big enterprises let's call them the big extra large enterprises you know they've a lot of them have tried things out some of them like in the dreamforce keynote pfizer will be there they've deployed agent force to 20 000 you know their sales professionals they use our life sciences cloud which is now also with Gentic.

25:06That idea that we're able to kind of, you know, work with a company like that to kind of give them much more capability for their professionals, that it's a force multiplier, you know, for their employees. That is very powerful. And for their customers as well. And other customers need to see that. In the pharmaceutical industry, it's a great example. You know, it's a variation of companies who are willing to try technology to companies like Pfizer, who are always the leaders in deploying that technology. So, you know, I'm called Albert Borel, CEO of Pfizer. Please come be part of my keynote. That's very important for me because they are pioneers.