Marc Benioff on AI Eating SaaS, San Francisco’s New Mayor & CEO Trump Pandering

[00:00:00] Tom joined the podcast and he got us Marc Benioff. You ask, I delivered. You write antagonistic stories about Microsoft and suddenly Marc Benioff wants to come on the podcast. That's not what this was about. This is about a spirited conversation about the future of AI and tech in San Francisco. We show we're really willing to hold Microsoft to account. It's like Salesforce is like, that's the kind of podcast I want to go on.

[00:00:26] Marc Benioff, Yeah. Look, Marc, you know, I've known Marc for a couple of years. I wrote about him at the Journal. I wrote a pretty tough piece on him actually. Marc Benioff, That's true. Yeah, I remember that. Marc Benioff, Didn't he throw like the Governor and Yo-Yo Ma at you? Is that in the story? Marc Benioff, Yeah, yeah. They did both speak on his behalf as a great guy. They're still on my phone, by the way. Marc Benioff, Future Guest by the way. Marc Benioff, It's like, oh, Tom might be doing a negative story about... Marc Benioff, Listen to the end of the podcast, by the way.

[00:00:55] Marc Benioff, I give my unadulterated take on what I think of Benioff, but Tom was working on what could have been a negative profile. What was the angle? Marc Benioff, At the time? I mean, it was a very tumultuous period for Salesforce. They had five activist investors in the stock. Marc Benioff, There were major concerns about their margin and the angle was like, oh, and they just laid people off.

[00:01:17] Marc Benioff, And so that was a story basically saying, for those of you who don't know, Salesforce has this famous Ohana family culture. And the story was like, what does it mean when your family lays you off? Marc Benioff, Yeah, that's a good question. Marc Benioff, Yeah. Marc Benioff, And so anyway, and it ended up being kind of a profile of Marc. Marc Benioff, Turns out business is just business sometimes. Marc Benioff, Yeah. And listen, to his credit, Marc is the owner of Time Magazine. He understands how media works. And I think he felt like he wanted to always...

[00:01:46] Marc Benioff, Make sure you knew that Yo-Yo Ma thought positively of him. Marc Benioff, Yeah. Although I encourage people to go back and read that story because the quote from Yo-Yo Ma is very funny. Marc is a technologist. I like talking to him because he came from that era of Silicon Valley.

[00:02:04] Marc Benioff, I think one of his first jobs was working for Apple and he knew Steve Jobs. And so for him to kind of come on and talk about AI as like along the same spectrum of importance as the internet and mobile and cloud is meaningful. Marc Benioff, Yeah. We asked him about internet waves. We asked about what Klarna basically saying, fuck SaaS, I can build it myself. We asked him, do billionaires have better AI than the rest of us?

[00:02:30] Marc Benioff, I gave him a hard time about why billionaires are sycophants to Trump and whether he would take a different tactic. You dug into San Francisco, which I'm a New Yorker. I barely understood. What else? Any other highlights? Stick around. Listen to it. It's good. It's Marc Benioff. Marc Benioff, Yeah. He's an entertaining guy. He's one of the few people in tech that can actually hold an audience. And I think I'm interested to hear people's thoughts on this one. Marc Benioff, Yeah. Yeah. It's a fun one. All right.

[00:03:00] Marc Benioff, Give it a listen. Marc Benioff, We're excited to have now on the Newcomer Podcast, Marc Benioff, CEO of Salesforce, which is I believe the third largest software company in the world? Marc Benioff, I hope second. Marc Benioff, Is it second now? Okay. Marc Benioff, Well, we're going to do 41 billion approximately or 40.9 billion this year. So when you look at pure revenue, I think it is like probably the second largest enterprise software company. Marc Benioff, Is it overtaking service now or who's who's a- Marc Benioff, They're 10 billion.

[00:03:28] Marc Benioff, They're one quarter a size. Marc Benioff, Okay. Okay. All right. Well, we'll have our research department look into this. Marc is also the owner of Time Magazine, which we'll talk about. Marc Benioff, I'll be your numbers, Tom. Marc Benioff, Yeah. We'll run it through one of the Agent Force programs and see where it comes out. Marc Benioff, Exactly. Let's get the AI on the podcast to get clarity on all of this. Marc Benioff, Yeah. Yeah. We'll see who's right. And Marc's also a long time technologist and San Franciscan. What are you? Third generation? Marc Benioff, Fourth.

[00:03:56] Marc Benioff, Fourth generation San Franciscan. So we're always got a lot to say whenever I chop it up with Marc about tech. Marc Benioff, You're like first. Marc Benioff, I was born in the East Bay. Marc Benioff, Oh, okay. Marc Benioff, Yeah. Marc Benioff, All right. Well, you can still come over with me. Marc Benioff, But my son is first generation San Franciscan. Marc Benioff, You can take Bart over. Marc Benioff, Exactly. Marc Benioff, We used to call it bridge and tunnel people. It's no longer appropriate. Marc Benioff, Trans Bay, Trans Bay Tube people. Marc Benioff, No, we called them bridge and tunnel people, people who came over on the bridge or the tunnel.

[00:04:23] Marc Benioff, Uh huh. Because you guys thought you were more sophisticated as San Franciscans. Marc Benioff, Yeah. Marc Benioff, Way more, way more. Marc Benioff, Yeah. Marc Benioff, We'll get into San Francisco. Marc Benioff, Okay. I want to start off with a serious AI question or just how you think about the artificial intelligence wave we're experiencing in the context of the history of the dot-com boom, then mobile, and then cloud. Marc Benioff, Like what similarities and what differences do you see? Marc Benioff, Yeah. Well, certainly it's not the first gold rush in San Francisco.

[00:04:53] Marc Benioff, So is it a gold rush? I guess that, yeah, that gets to it right there. Marc Benioff, Well, there's going to be a lot of gold. So, yeah, I would say it's a gold rush. I would say that, look, it's a city and an industry marked by a history of innovation. And you've, you named a few of them and there's many others. And this one is incredible what's happening. It's obviously going incredibly fast because a lot of the technology is not proprietary. It's being built in an open environment.

[00:05:22] Marc Benioff, So is it a good source? Marc Benioff, It's shared. It's more of a communal asset for the whole industry. That's new. Usually these things tend to be very proprietary. And so they go slower. You know, like my iOS or Android, like, you know, these operating systems, mobile, you know, phone operating systems, like, like as a category, it's going to go slower than AI because everything's more proprietary.

[00:05:52] Marc Benioff, If one company has to build it slow, if a whole industry is racing towards it, it will be very quick. Marc Benioff, Exactly. Together and shared. Marc Benioff, Do you see, you know, what, like, you know, mobile had the app store cloud had software as a service, like, there is some more excitement about sort of usage based pricing in the AI world? Or how much do you see a sort of a business model transition along with like a technological one?

[00:06:19] Marc Benioff, Well, I think that there's going to be a business model transition, a technology model transition. There can be usage based pricing. There's also per user price pricing, you know, it's all encompassing because it's connecting to every aspect of the technology. I'm sure you know, like, for example, my chat TPT, I'm paying by the user per month right now. It's not usage as far as I know. That is, you know, I'm paying like 20 bucks or something.

[00:06:49] Marc Benioff, Chat TPT or Claude. I have, I have Grok. That's also per user pricing. Marc Benioff, Do Salesforce employees have access to one of the chat bots officially? Marc Benioff, We have probably many. Marc Benioff, Yeah. Marc Benioff, Is my guess. Marc Benioff, There's not one. There's not one. Marc Benioff, We're investors also in a number of these companies. We also deploy a lot of them directly through Slack. And then, you know, Slack is becoming kind of this meta level for AI, where I'm talking to the AI, the AI agents are talking back to me.

[00:07:19] Marc Benioff, I'm sharing, I'm collaborating. They're talking to each other. Marc Benioff, But explain to me if you want to put it at the same significance of cloud, of mobile, of the internet itself. What is the transformation exactly here? Like, explain it to me as if I was, you know, a mother in the Midwest. Marc Benioff, A golden retriever. You know, I feel like everybody's rewatching that movie. Anyway, go ahead.

[00:07:40] Marc Benioff, It's a good question because, well, yeah, movies. Let's take movies. We saw War Games. We saw Minority Report. We saw Terminator. We saw Her. And we realized that in all of these movies about the future and the sci-fi, there's some AI in there. There's some robotics in there. There's some drones in there. Seems like it's all happening simultaneously.

[00:08:00] Marc Benioff, That said, here I am in my office and there's no robot walking around yet. Here we are on this podcast. We're on the newcomer podcast. There's no AI that we're talking to on the podcast today.

[00:08:13] Marc Benioff, So we're not quite there yet, are we? So not to say that maybe in a year that might not be true. Like I'm waiting to, you know, we have you guys. I'm sure you're great, talented, well-researched. But I think we can see that there's the potential for a level of scale with this technology that's amazing and our imaginations are all going wild and are excited.

[00:08:38] You know, how's it going to change my, you know, inter-side software industry. And, you know, we're getting ready to report our earnings, you know, in a few weeks and we still have to manage our financials and manage our HR. And I thought you'd be sort of boom time AI. Weren't you shouting agent force from the rooftops? You sound a little bit more measured or have you seen things that make you more apprehensive about what's going on in AI? Well, I think that I want to get to the other side of the coin.

[00:09:08] The other side of the coin is I would say for the last 26 years at Salesforce, I'm working in this multi-hundred billion dollar enterprise software industry. That's like our TAM or we call our total addressable market, you know, building software that addresses all of those things. Today, I walk you through these customer examples where we're building digital labor for these companies. And that's a maybe three to twelve trillion dollar opportunity. It wasn't that long ago, maybe six months ago.

[00:09:38] Tom was still employed at the Wall Street Journal. He came to Dreamforce. He sat with our customers. He was building agents. Multiple times. It took you a while. It took you a while to get your organization together, to have me sit with a real customer. I recall that. I watched him build an agent and deploy an agent. And maybe that was the first time he built and deployed an agent with Agent Force. And, you know, you know, you never forget your first.

[00:10:03] So I'll just say that this is like an incredible moment where, yeah, there's huge potential. It's much bigger than ever before, but we're at the beginning of our journey. So how do we think about that? It's very simple. We always overestimate what you can do in a year in our industry and underestimate what you can do in a decade. And certainly underestimate two decades. And in the case of Salesforce, underestimate what you can do in two and a half decades.

[00:10:33] So, yeah, it's an incredible moment. And we should all feel completely excited that we are alive to watch this and to work with these large language models to become more productive, to augment ourselves, to deliver digital labor, to improve our society in many incredible ways like health care and education. It's awesome. How much has AI transformed your life?

[00:11:00] I mean, one thing that's interesting about the iPhone, right, is that sort of the iPhone I have, and I presume whatever iPhone or phone you have are the equivalent. There's no sort of like billionaire iPhone. Like with ChatGPT, with sort of AI models, I mean, you could build your own open source stack. Like do you have sort of a better AI than we do?

[00:11:24] And how much like custom stuff are you trying to put together to sort of see if it can come up with better ideas for like Salesforce than you do with your team? A day does not go by that I'm not using one model or another. I would say that, of course, Salesforce, we've been fortunate to be pioneering many of these AI tools for many years. Everyone knows prompt engineering really came out of Salesforce's research team. We did the first coding models. Everyone knows CodeGen.

[00:11:53] We have other models that we, XGen, others that we've done that have been pioneering. We felt like our obligation was primarily to do the research and contribute to the open source community and then build it into our core platform. That has been really our mission around AI. When it comes down to myself, rarely does a day go by that I'm not using one of these amazing AI models. I use it in all kinds of different ways. Right now, we just finished our quarter. I have all of our numbers.

[00:12:22] I'm looking at our numbers. I'm, you know, using models to kind of look at my business plan against my performance. Look at what's working, what's not working, how to tune, how to improve, how to refactor, how to restructure. Before I had to use an analyst. I don't know. Are your analysts trying to tell you, no, you're getting the numbers wrong. You need humans like the hallucinations. I am the analyst. I am the analyst. And I'm the individual contributor. I'm a manager. I'm a leader.

[00:12:51] I do all three things. And as an individual contributor, I'm much more productive than ever before because I have tools that are giving me the ability in many cases to be unlimited. So I can come up with key insights about, you know, not only just my life, but my company, my industry. And by, yeah, it's also very important to learn how to ask the right questions. None of this happens without the right prompt. I don't have to tell you that. You probably know. Like, it's the one thing.

[00:13:19] Like, it's not going to come up with all of those questions. I sit now with a lot of people, by the way, different kinds of people. Like, I have a friend of mine just lost his job at the FDA, took the retirement package. He's thinking about what he wants to do next. He's like a real technical expert. He's like telling me how much he knows about the AI. I sat with him and I'm like, hey, let's just ask it a bunch of questions based on what you know about the technology.

[00:13:44] And we kind of ferreted through to actually build a business plan, a slide deck, a company name, a logo, you know, all of these. He couldn't believe it. You know, he thought he had had it all wired up. And then he came out of it super motivated where it's like, whoa, I can build this next generation. To this specific, just do you think you have a fancy – do you have access to a fancier system than we do? Or do you think – or you're fundamentally using sort of the mass market and you see an equivalent?

[00:14:12] I think that I just want to use the stuff that everybody else is using. But, you know, I'm an investor personally in some of these things like U.com, Richard Socher, who's been an incredible AI researcher, a lot of folks know, came out of Stanford, deep learning pioneer, was the one who I think authored the paper on prompt engineering. He started U.com. That's a company I'm personally invested in.

[00:14:35] And Salesforce has invested in a number of the really great model companies like Anthropic and others, Coher, others, Mistral, Hugging Face, not OpenAI because only one company was invested in them. Which company was that, Mark? Starts with them. Yeah. Yeah. We might get to them. I actually want to go – you mentioned my demo that I had at Dreamforce last year. And, you know, that was through Slack. We built an agent.

[00:15:03] I'd rather call it an initiation than a demo. Like a gang initiation? Essentially? I won't – I'm not going to have more crass terms that I could use. Okay. Well, whatever you want to call it, I got to build – We were aggressive with you, I think. I had to personally bring you down and say, Tom, we're going to now build an agent. This is going to be important for your career to understand what agents are. We're going to show you how these models work, how the data works. Let's just do it piece by piece. Come on.

[00:15:30] I got the applesauce out and the little tiny spoon gave you a little bit. You're like, that's so good. Another little spoonful, another little spoonful. And then you were like, oh, this is so yummy. And I said, yes. Yeah. Yeah. That's, by the way, every conversation I've ever had with an editor in my career is a downstream version of that. So you see – now you see what my life is like. And I imagine your customers too. But that was Agent Force original.

[00:15:58] That was basically building a prompt interactive agent. And then a couple of months later, I get invited, end of the year, to another event, this time held at a hotel, where you guys are rolling out Agent Force 2.0, which is agentic. These are agentic tools. So within a matter of months, you guys went from rolling out a basic chat bot type. Well, Agent Force was agentic at the beginning, Tom. And it was also – but what version 2 had was the reasoning model.

[00:16:26] That was where we really had broken through and we started to really say, like, it's not just LLMs that are important, but also reasoning models and LLMs together are, like, going to provide us a new level of intelligence and capability. That's where I think a lot of us should be more excited than we are because, you know, LLMs are cool and have come a long way, especially in the last, like, four or five years. But reasoning models are, like, pretty brand new.

[00:16:51] But the way that they learn is not with these kind of somewhat semi-static data sets extended through these RAG techniques, but they're, like, stacking their intelligence to expand their capabilities. And I think that is something that's still somewhat underappreciated, that LLMs plus reasoning models are going to give us some kind of – we're still in early days, but it's a case for optimism. But also a case for complication in the marketplace, right?

[00:17:20] Because you guys are – you just introduced, you know, and you guys have a lot of enterprise customers, you know, not digitally native customers that are using – have been using Salesforce for years. They just learned about, you know, Agent Force 1.0. And then a couple of months later, you guys come and explain agentic tools to them. Why is there so much change in the technology business? But it's a tough sell. It's a tough sell. I know because, you know, Mark, I also talk to customers, and I know it's a tough sell. Okay. This is what you're trying to say. Yeah.

[00:17:48] The rate of innovation is exceeding the rate of adoption because we are going so fucking fast. Like, it is really fast. So the rate of innovation is, at a moment in our industry, exceeding even what customers can accept and adopt. And some customers can. Like, I have all kinds of stories we can talk about. Well, you know, you saw thousands of customers being onboarded into Agent Force in September and October.

[00:18:18] It's only been six months. Now we see so many of them with deployed systems. That's amazing. But there's – you know, we have hundreds of thousands of customers. So we don't have hundreds of thousands of customers on Agent Force yet. We have hundreds of thousands of customers on social or on mobile or on the cloud or, you know, on analytics or on Slack. We have, you know, almost a million customers. But, you know, we're still at the beginning of Agent Force with single-digit thousands.

[00:18:47] And those customers are just, you know, at the beginning of their implementations and their tuning and setting up – remember, Tom, we set up the guardrails. And we set up the semantic layer. And we're setting up the action layer. And we're setting up the application itself and all the structures so that we can kind of make this work for the customer. And so, yeah, we're still at the beginning of an incredible journey. And we should all be pretty jacked right now in our industry because our whole industry just kind of boomed, just got reborn.

[00:19:15] And at some level, you know, I just heard this last week, so I'll steal it, where humanity is kind of like a chrysalis, like giving birth to this new form of life, which is this artificial intelligence that's going to be this kind of self-learning, self-growing technology. And that is kind of amazing what's happening. Like, I can't even completely get my head totally around it. We've seen it in the movies. Are you falling in love with your AI bots over here?

[00:19:44] You're imputing consciousness onto them? It becomes hard to sort of not sort of get attached to your chat history the more context it has. Yeah, well, I mean, it's going to be a different kind of consciousness. I mean, like one thing is, you know, like us here, you know, especially, you know, for those of us who have, you know, you know, we know what suffering is. So therefore, we have compassion because we know suffering.

[00:20:12] So AI doesn't know suffering exactly. So it can simulate compassion or simulate empathy, but it's not really compassionate because it isn't suffering. So, but some may. So like I just saw like on Twitter yesterday, there's a brain organoid, you know, which is kind of we can talk about that they've grown an organic AI that's powering a robot. And when you look at that, it's on my Twitter feed. It's like crazy stuff.

[00:20:40] And then it's like that type of AI may actually be suffering. Like it may, it's come out of brain stem cells. They're, they grow these organoids. Whoa, that's like, yeah, that's like a different kind of thing. We're talking about in organic, you know, AI that's growing as an organism. Yeah, but these are the things you guys are, I'm assuming, talking about on the front lines with your sales team trying to get enterprises to buy Agent Force 2.0.

[00:21:08] I'm not selling any brain organoids to our enterprise customers right now. That's what you're asking. You're not there yet. I'm just out there riffing. But no, if you want to know what I'm selling to customers, like we have a customer smart sheet, which you probably like. It's like a cloud spread, like cloud Excel, you know, and they've got 13 million users. They're an amazing company. And they've built Agent Force right into the app. So you're like working with it. You can get your support and you can get your service. You can increase and decrease the number of users, your license force. You can change your password.

[00:21:38] You can do all these things on Agent Force right inside Smartsheet. You can go to 1-800-ACCOUNTAN-NOW to do your taxes. And 70% of all of their questions have been fully autonomously answered by Agent Force now. So that's, you know, has a high cool factor. So let me cut you off there for a second because this is interesting to me because I sort of see two implementations of what you guys are selling.

[00:22:05] There's the internal workplace stuff, you know, ways to make a workforce more efficient, the way you interact with your coworkers, putting together reports, slideshows, whatever. You know, writing up marketing documents. And then there's customer-facing stuff. Explain to me the progress you're seeing on customer-facing stuff. Yeah. So it's bifurcated like you just said. There's two worlds. One is we're augmenting ourself with AI. We all know that. I'm on ChatGPT, Anthropic, U.com, Grok, Gemini, whatever.

[00:22:34] And I'm learning. I'm growing. I'm asking it questions. Hey, what do you know? Tell me everything you know about me based on my chats. You know, hey, look at these numbers. Tell me, you know, how do I make this division even more successful? And then there's the digital labor. So that's the productivity component. So we're deploying service agents or sales agents or marketing agents or, you know, sales development representatives. And these are autonomously working.

[00:23:04] Like I just said, like for 1-800-ACCOUNTANT, we provided them digital labor that resolved 70% of all of those questions that they had. Or if you go to help.salesforce.com, the one I have, we've done more than 500,000 customer resolutions, you know, using AgentForce. So it's a split. Like over here, we're augmenting human beings, making them more productive.

[00:23:28] Over here, we're creating new kinds of digital labor, giving a company a new kind of employee. And an example like with Gucci, when we gave them this technology and we said, hey, we're giving you digital labor, they actually transformed it and used it only for human augmentation and raised their sales 35% in their Florence call center. That was an amazing example. And I'm sorry, how was generative AI able to raise sales by 35%?

[00:23:57] It's basically an AI interacting with a human being. You're saying an AI did a better job. So Gucci has this amazing call center in Florence. We took our technology. We delivered it to them. How are they going to use it? We kind of just showed you two ways. And these employees only were doing, you know, customer service. Like, hey, I have my bag I need repaired. Oh, yeah, we'll repair your bag. Hey, I bought this garment from you. I've torn it. I need it refined. We'll send it in. We're going to send it back to you.

[00:24:27] It's just a customer service center. Then, all of a sudden, those people were able to really understand all the products. They're able to understand all the capabilities. So they were able to sell, not just service. And they turned themselves into a sales and service center. And revenue went up 35% in that Florence call center. So that's an example. I'll give you another really cool example. Disney.

[00:24:53] So Disney, you know, we've been running all the customer touch points for years. So, like, the Disney store, you know, the people selling the cruise ships and the real estate and the, you know, the cast members. They call them cast members. These are their employees. Like, in the park, they have the Disney guides. You probably, I don't know if you have kids, and they take your, you know, their Disney guide gives you, like, a tour around the park.

[00:25:16] And, you know, on those phones have always been Salesforce and Slack, where they keep track of the tour and then talking to other guides. Now, you know, the agent force says, hey, you're bringing Mark to Galaxy Edge. You're bringing him to the Millennium Falcon ride. But it just, it's down right now. So I know you're on your way. I can see you're halfway there. You need to make a right turn and take him to Toontown. I just looked at all of Mark's rides, everything he loves, everything he's looked at on Disney+.

[00:25:46] Mark has opted into all of this, but that ride is down, but he's going to love this ride over here. And before, that guide would have to kind of stop and go, oh, too bad. Millennium Falcon is down. What do I do? Talk to his friend. Oh, hey, have you thought about taking Mark over here, maybe over here? In this case, the agents already looked at every ride, all the lines, what's working, what's not working, my ride history, the weather patterns,

[00:26:13] every single piece of data point it needs to make the right decision and then say, nope, Toontown, that's the one for you. Right. So essentially what you're describing here is a tool that assists an employee to make them better at their job. A co-pilot, if you will. Augmented. Augmentation. Augmentation. Fundamental augmentation. Sure, sure. Some people call those co-pilots. But that doesn't sound to me like a digital labor force, though. You're talking about digital labor. That's not replacing someone.

[00:26:40] That maybe makes them better at their job, maybe makes them more profitable for their employer, that they get greater revenue out of their interactions with customers. But how is that digital labor? So digital labor is two forms. One is that, you know, like right now, look, I have 9,000 support agents at Salesforce. Okay. And we have this new agent force also. So we have a service force. Now we have an agent force.

[00:27:09] And that agent force is resolving a lot of customer support and service issues, even doing some sales autonomously without interactions from humans. So those humans have actually quite a bit less to do because it's being done by the digital labor. And another example is there are robots that are made by all kinds of different companies today. And some of those robots are just manifestations of agents.

[00:27:38] And those robots could be on assembly lines. They could be in your home. They could be in a hotel cleaning. And they are interoperating and interacting with those, you know, data sets. Is Salesforce in that business? Like, are you going to try to build the agents that might power a robot? Or is that an area that... We've already done it. We've demonstrated it in our last three conferences that agents get manifested into these robots, as I think is a good way to say it.

[00:28:07] Would you say that, you know, here's these incredible robots, amazing technology. But the robot has a model that it's operating on. The model's kind of helping it do this, move around. Like customer support. Customer support, clearly number one in terms of where agents are happening, right? If you look at like the Aloha model that's come out of Stanford, which is like a really good model for training a robot. And like the videos are mostly these robots cleaning these hotel rooms.

[00:28:37] And so there are these cool robots. They're cleaning their hotel rooms. They're trained using this Aloha model. And now I open my door and I come in. There's a robot cleaning my hotel room. And it says, oh, hi, Mark. Do you want us to leave the room? And, you know, thank you for being a good customer of our hotel chain. And how do you want it? Sorry, we're here. Do you want us to stay? Do you want us to leave? And how do you want us to interoperate with you in the future?

[00:29:07] And is there anything else you want us to say to management right now? Or is there anything else that you need? So there is an interoperability between that digital labor, the robot, the agent, and even what we do, which is the customer platform. You think white collar jobs are the most threatened still? Or you think this robot use case opens the possibility that we see sort of aggressive progress in blue collar jobs? Well, we already know there's robots in the factories.

[00:29:36] And there's, you know, there's, I mean, robots are a part of life already. We just maybe not have them necessarily in our homes. Well, maybe we have, you know, some of the vacuum cleaner or lawnmower robots or something. But I have that Unitree dog, you know. And it's hooked up to Salesforce, by the way. So it can see, it can talk, and it can interact with Salesforce and talk back. But it's somewhat still limited functionality.

[00:30:02] But when you look at some of the things that are coming, whether it's Tesla Optimus or Figure or a lot of these new robot companies, Unitree, et cetera, and you start to have the ability to operate in a home or business environment and connected with an agentic platform, it's going to have a lot of functionality. That said, here I am. Let me just check. No robot here. No, we're not there yet. And look, you know, what we've been talking about so far seems to be a very rising tide situation.

[00:30:31] You pitched, you know, the Salesforce product in particular. I tried to trigger you earlier by bringing up Copilot. You, Tom, tried to trigger me? Yeah. Mike, you would never do such a thing. Well, let the audience decide if I got a rise out of you. But, you know, I heard an interview with this guy, Satya Nadella, who basically described AI as a replacement for SaaS. That effectively, you know, these CRUD databases, which is effectively what SaaS is. Yes, right. All right.

[00:31:01] Give me your response here. And that's what he's done at his company. And Klarna is saying it, right? And Shopify is using it. Right, of course. Yeah. No, no, let's not get distracted with Klarna. No, no, no. Microsoft, Microsoft. Sorry, sorry. They would true their claim. Those guys would true their claim. Yes. Right. After pressure from people like you, by the way. I'll give you credit for them. I mean, look, let me give you my vision, which is first we just described digital labor revolution, which could be agentic. It could be robotic. These are two different types of things that are happening that are so significant.

[00:31:30] And there's many other things in the industry. We can talk about satellites and comms. And, you know, there's many things happening all at once. But for sure, AI is the biggest transformation. Now, when we talk about the fundamental transformation of software and is software going to change and transform, there's no question it's going to become more dynamic. It's going to become more intelligent. It's going to mold itself directly to me, all of these things. But, you know, we also know what reality is today.

[00:31:59] And just, you know, I have my phone here. I look down at my phone. And let me just say, transform my phone into the Mark Benioff phone and only show me exactly an app designed for me based on my preferences and the user interface and the type, you know, for my age that's appropriate for me and size and language. Oh, wait, it didn't happen.

[00:32:24] But you could imagine that there's a moment where the software will become more dynamic shaping to me. And we have that running at Salesforce also. But let's also talk about what Salesforce is. So, you know, it's a series of applications like sales, service, marketing, commerce, Mule Software integration platform, our, you know, Tableau, Analytics, Slack. You know, I think you've used that now. We are Slack customers. Slack, thank you.

[00:32:53] And we weren't at the journal. So I had to leave so I could join Slack. Thank God. And then here's the thing that's interesting. So now we also have a data cloud that's deeply integrated with all of those apps. And we have our agentic layer that we just went through. So now we have a new version of Tableau. Tableau is the most popular business intelligence analytics platform in the world. Incredible company. Acquired it, I don't know, seven, eight or nine years ago. I can't even remember how many years ago.

[00:33:19] And we've rebuilt it completely and just showed the new version two weeks ago at the user conference in San Diego. And it's the Slack that all these customers love. They call themselves the data fam because they're like the data family. And but we put on an agentic layer, agent force. And we also hooked it into our data cloud so that they could all work off of common data repositories, but then integrate with other things. It's also embedded inside Slack. It's also embedded inside all of our applications.

[00:33:47] It also is a marketplace of all these other applications. It has an action layer. It has a semantic layer. I get the connectivity argument, but, you know, aren't you a little worried that like, you know, I could take a picture of Slack. I can go in like lovable or cursor and it will rip off sort of your general design and some of your features. Go for it. And let's see how far you can get. And by the way, you might be able to get some distance, but I think we know there are limits right now.

[00:34:16] Prototype, I think, you know, like first draft. And I think eventually, yeah, maybe all software will be built and designed by the AI. We won't need software companies. But today, I think you'd really love Tableau next. If you're an analyst, you can sit there, you can talk to it, you can work with the model just like you're doing. You can do all those things and you can get all the visualizations. And you're still going to need high quality data management with a sharing model and a security model.

[00:34:43] Look, if you're at a bank, you can see your customers' account balances. You cannot see my account balances. You know, like there is how data has to be highly regulated. It's managed. It's governed. Security is critical. And then you have those three layers, the application layer, the data layer, and the agentic layer. So then we can start to decide, all right, which of these layers is transforming? Which one's going away?

[00:35:08] What about the foundation, the fundamental platform layer as well, which might include things like the federation of the data, the search of the data? This is the life of a technology CEO with journalists. At first, our line of attack is, oh, you're coming up with technology too quickly. Your customers can't figure it out. And then we're like, but your customers are going to rebuild your technology from the ground up and they're going to be as sophisticated as you. By the way, that's just why I love what I do. This is why like 26 years later, you know, here I am.

[00:35:37] I love this and I just want to, you know, let's just ride this wave and like go with it. You know, this is going to be fun. It's going to be fun for my customers and it's going to be fun for us. And we want to take you along with us and we want to go with you and we want you to contribute. We want to contribute. And I was teasing Tom earlier about how, you know, we showed him what an agent and agentic layer is for the first time. That's meaningful to me. Like we want to be ahead. We want to show you.

[00:36:04] We want to be like, we want to be, hey, the first time you ever heard about agents and agentic was Salesforce, you know? And we have some new stuff coming that you're going to get that feeling from as well. Do you use the product? And I want to go to San Francisco. So, but I can't help. But, you know, we use Slack and we use AI all the time. Like I'm constantly on chat with GPT. Why isn't Slack sort of harassing me a little bit more, you know, to do stuff with AI in the platform? Or like human, sorry, superhuman.

[00:36:32] My email client, you know, almost my chagrin is sort of trying to force its AI features on me. How much do you think about you're basically saying why is Slack not more like Clippy? Or, yeah, should you be? Or like how do you think about the tradeoffs of like you like Slack. We're not going to like clutter it. Slack has been and I think that Slack has really evolved beautifully and has really embraced AI aggressively. And, you know, there is another level and you'll see more functionality getting added.

[00:36:58] And you're not going to are you going to force like how much do you see across the company pushing it? There could be a nudge. Yeah, yeah. Here's my last question on this topic here and looking just broadly across tech and, you know, the years that you've worked within it. How disruptive is AI in terms of the order of the balance of power within the industry? Is there an upstart that you see that could actually upend and take down a large incumbent right now, making them obsolete in the way that we've seen in other revolutions?

[00:37:27] Because as I see it as a reporter, it seems like almost aside from potentially what open AI becomes, it's all been owned by the tech giants. All the transformation, all the upheaval, all of the disruption is basically been owned by the cloud companies, the big software companies like you. And I haven't seen the kind of, you know, what is the Uber of AI? What is the, you know, the AWS of AI? Well, I don't know. I think you are seeing, first of all, with AI, we have to remember open source is a key part of the revolution.

[00:37:57] A lot of this innovation has come from open source. So let's bow to open source because that, without that, we would not have the speed that we have. Number two, you are seeing a lot of new companies. I mentioned a few of them like OpenAI, Anthropic, U.com, Perplexity. I mean, there's a list of them and many other hundreds, thousands of other AI startups.

[00:38:25] And then you also have, yes, the established companies as well who are also trying to do that. There is some tension. You can see it even with the Microsoft OpenAI Alliance. So you can see that Microsoft is working on their own large language model. They have hired Mustafa Solomon, who was not just the CEO of inflection, but before that was the founder of DeepMind, incredible executive known and myself for a long time. And there's no love lost between Sam and Mustafa. We all know that.

[00:38:55] And I've seen it myself firsthand. I can speak to it with authority. And so all of a sudden you see OpenAI, who was deeply dedicated to the Microsoft infrastructure, saying that they're going to build their own independent data centers and are building the first one, Stargate, you know, with Crusoe and, you know, their CEO, Chase. And he's on the case.

[00:39:19] So I think that this is really like interesting that there is some drama there. And even though they say there's no drama, we know there is drama. And Tom reported the co-pilot, the co-pilot figures. Were you surprised it was as small as it was? Well, Tom knows I wasn't surprised. No, that doesn't mean Mark was a source on it. But certainly, you know, look, you've been on this for a while, Mark.

[00:39:48] I was not surprised because I've been saying for a while that, I mean, I use the products. How hard is this to figure out? I have these things on my phone. It's like this one doesn't work. This one does. They have GitHub. They have this kind of what they call, what did you call it? Co-pilot, Tom? You know, where the developer, and I thought that was the one that was actually pretty cool, where the developer had a co-pilot helping them write more code. You know, we had code gen we're trying to do with LMs. And then all of a sudden there's this company, Cursor, that comes along.

[00:40:18] And then there's Windsurf. And then it's like, and what was that over there? Oh, wait. Oh, well, so much for that idea. So it's like it kind of goes to show how you got to go faster. You have to be more creative. You have to, we're in a tornado. You know, we're, if you know, like my friend Jeffrey Moore, like he's written about this. We're crossing the chasm. There's like craziness underway. And yeah, each company is trying to do what they're doing.

[00:40:47] Like we're, we're doing, we're agent force first. We're building agent force companies. Like I already gave you some examples. Now I can give you, you know, more examples of the Disney. And I gave you 1-800-accountant. We want to make sure we talk about another passion of yours. San Francisco. I'm curious, is it a new dawn in San Francisco? Or how do you, how do you rate the mayor? How different do you think it is? I've known the mayor for a long time. And I'm excited. Look, we need to change. Everybody knows that.

[00:41:17] Look, I'm always an optimistic person. San Francisco has so many great things about it. I could go through it in detail. And look, I love San Francisco. I've invested a huge amount in the city, not just for my company, but, you know, large individual philanthropists. Built the children's hospitals in San Francisco and Oakland. Cancer research centers. Microbiome research centers. Other things that, regenerative biology.

[00:41:45] Other things that I'm, you know, really excited about at UCSF. We've put $150 million into the San Francisco and Oakland public schools. We've done a huge development of public parks. You've probably been to the Tunnel Tops Park and the Presidio. I think it is a truly an incredible moment for San Francisco.

[00:42:03] I also look at all the other things that we have tried to do with, you know, obviously we have had a huge homeless problem for decades since the 50s when my grandfather was a supervisor in San Francisco. I'm in these public schools. I don't know how many people are doing the research and are in the public schools doing the work. You know, public education is super important to me. It has been for a long time. I'm in the schools.

[00:42:30] A huge percentage of the kids in the San Francisco and Oakland public schools don't have homes. And these family, their family homelessness is a huge issue. And that really tugged at my heartstrings. I'd walk with my grandfather to work for sure when I was a kid. And, you know, the homeless shelters used to be on Market Street, right where the Twitter headquarters are. So that's not a new homeless area. The Tenderloin is not new.

[00:42:57] It's been there for four, five, six, seven, eight decades. And it used to be called Skid Row. It was like really like these are difficult things. And, you know, I have invested in and worked with maybe every homeless organization in San Francisco from the Salvation Army, you know, to UCSF, to Hamilton, to Glide.

[00:43:23] Each and every one, like if they're working on homelessness, they know who to come and ask help from. And Prop C also is about tens of thousands of families. I think there's a lot of misinformation on Prop C. So I was just with a friend of mine last week. We were debating it. And I'm like, let's get out ChatGPT. And we're going to ask it. For a Microsoft hater. I mean, I guess you're saying ChatGPT is distinct from Microsoft, but you do use ChatGPT. I use ChatGPT. I use them all.

[00:43:52] And then I was like the analysis of Prop C, our city, our home. And here's the data set and connected it to SFGov. All the data is there. And, you know, there's no perfect solution for homelessness. There's no finish line. You know, our society, capitalism, we have no guardrails. So there can be some extremeness. And we know we don't have enough homes.

[00:44:16] We know San Francisco is a city of elite folks who pay very high prices to live there. And a lot of people have to end up moving because of that. We are only on seven square miles for an incredible tiny sliver of land. And we're trying to jam all these people in there. And we're not like other countries where we build social housing right inside our mainline housing, which a lot of other folks do. So there's different ways.

[00:44:43] And homelessness has been something that I've definitely continue to think that we have to focus on. And I hope that the mayor continues to focus on that as a major issue. He ran one of the largest philanthropic organizations in San Francisco and worked with them a lot on all of these issues. Yeah. But let's also cut to the other part of this point, which is that when you came out in favor of Prop C,

[00:45:10] you really ran against the current of a lot of people in the tech industry and people like Jack Dorsey. And I saw even Elon, who you're fairly friendly with, essentially blaming you for the state of the city and that your support of this thing was the reason San Francisco is... Yeah. Well, it's silliness. And that's why I just went right to the AI and said, hey, you tell me. That was your unbiased response to Elon?

[00:45:37] That's why you got to go to Brock and send it to Elon. These guys are great salespeople. And you can say whatever you want on social media. But the cool thing about the AI and what I love about it, and I love that at Brock feature, is you can just say, hey, at Brock, tell me everything about Prop C, our city, our home, the results. Here's the data set. You can give it the URL, you know, from sfgov. And now tell me the success and failure of this.

[00:46:06] And is this all true or is this false? And you'll see, like, tens of thousands of kids and families have gotten homes. And that may not be the political orientation of some people. But housing and public education and public health are my areas that I worry about the safety net in a city like San Francisco. And I've invested in the safety net.

[00:46:33] And so, yes, public housing, public health, and public education are three things that I've invested in. And some of the other executives that you've mentioned, I won't go through their philanthropic history because I am friendly with them. I try not to call people out and assassinate people, you know, online. Look, they haven't done anything. This isn't what they do. They build companies. They move on. So that's not what I'm doing, am I? You can see my – you know my history.

[00:47:01] Well, you also remain in San Francisco, which a lot of these people didn't. They're either, you know, in Austin or in the case of Dorsey, Africa. Would the world be a better place right now if you had bought Twitter versus Elon? I don't know. I would – I can't even imagine. By the way, when I – Did you get close? When I – yes, but when I was looking at that, my dreams and my vision for that wasn't any of these things. And then I don't know. The political environment has become so hot.

[00:47:29] I don't know if I would have been the right owner. So, I mean, I think that he's been a much better owner. I mean, it's been impressive to see what he's done with the platform. I think – Well, he hasn't made it a great business. I mean, maybe he's potentially made it into a more controversial – I don't know the numbers. I can't speak to that. Yeah. I mean, you're – you know, looking nationally. I mean, you're sort of an outspoken person who sort of leads with your heart.

[00:47:52] It feels like we're in this moment where tech executives are just sort of cozying up to Trump because that's sort of the smartest move. Or what do you make of some of your peers just sort of like flattering the president? You weren't there at Inauguration Day, were you? No. I mean, let me just speak to this directly. So, number one, I've worked inside a Republican administration. I was Republican. I worked inside the Bush administration.

[00:48:20] I was the chairman of the PCAST when I was in my 30s. I realized that I'm better as an advisor and outsider and consultant, and I decided I would never take another government job after that. So, that was where I am now. It's one of the reasons I bought Time Magazine because I could be independent. I wasn't going to make political contributions. I haven't made any political contributions of any type since 2018.

[00:48:47] So, that is my highest order bit around that. And I would say that when we want to talk about, you know, presidents change, my values don't change, and then I try to align with these presidents. And Trump is no exception. I've aligned with him. If you remember in the first Trump administration, I came up, we talked about homelessness. I also focus a lot on forests and reforestation, which I believe is a major issue in the climate.

[00:49:18] And Trump aligned with me, and we launched one trillion trees together, you know, and it's been written about extensively, but he got completely behind 1T.org. And by the way, that's been an incredible thing, even globally, where even China has come in with 70 billion trees, and Xi Shenhua, you know, came forward and said that they're joining. And the Republic of Congo has joined, and many people have joined.

[00:49:42] And so, that's an area where I can align, and I hope that I will find other areas. I already know there's going to be other areas. But you don't think CEOs are uniquely afraid to criticize Trump? I mean, it is a different mood than in the past. I think that you have to find opportunities to align and move forward. And that's what I tried to do on the first one.

[00:50:06] By the way, what I learned in the Bush administration was that I was not going to agree with everything in that administration. I won't go through in details. Obviously, one of my closest friends was the Secretary of State, Colin Powell, who ended up joining our board. And I just learned that I'm not going to be able to align with every aspect of a political organization because I have my own unique values and beliefs and vision for the world.

[00:50:36] So, but I can – I don't have to do everything, but I can align on some things. And I think that if I don't align on those some things, I'm doing myself a disservice to my own core values. And that's where I found, for example, the opportunity to work on reforestation. And I will find other – I guarantee you now because I already know I'm working on it – other things where I feel good, where I can say, yes, these are my values.

[00:51:01] And I'm going to – you know I'm doing like huge programs on ocean philanthropy and I'm very worried about ocean health. I'm worried about these same things. I will find the opportunities. And by the way, I've tried that, for example, in the Biden administration and I only – I made some progress on some things and other things I did not. Like, for example, the Replant Act I felt was very important. And then other things I was very disappointed on ocean, how they assessed and moved forward on the oceans.

[00:51:32] Is diversity, equity, inclusion dead at Salesforce still alive or where – I mean this administration hates sort of DEI. Are you holding the line there or how do you think about that? I've always just – I'll just tell you how I look at it, which is I've always looked at that totally differently. I've really believed that it's about, you know, do we believe in equality?

[00:51:54] You know, I believe that we're all created in some level with equality in our hearts and that we can come together on this. And so we should provide equal opportunity. We should provide equal access. And we should provide equal pay. And that means that specifically like on men and women, you know, we audit at Salesforce. Do we pay men and women equally for equal work? And a lot of companies don't do that. Just – it's an easy audit.

[00:52:22] We use this thing called an HR management system. And we can just say, do we pay – we could ask the LLM, which we do. Do we pay men and women equally? And it'll say, yes, here, here, and here, but not in this country or in this acquisition you just made. And then we'll equal it out and kind of feel good about that. And then because of that, we have, you know, a lot of operating units that are female-led.

[00:52:46] That's – nothing that we've actually targeted has happened like that marketing, legal, you know, finance, communications have become female-led organizations. Some are still dominated by men. That's great. Like engineering others is fine. And then, you know what? They're shifting over time. They move around.

[00:53:09] And because we're not assigning those, we're not creating quotas, we're, you know, just saying let's just let this – But you did previously. I mean DEI was about creating, if not necessarily a quota, about an emphasis on hiring people that were underrepresented in these companies. I mean you guys have gone through a transformation in which you previously did include diversity as a criteria, and I think just this year took that out. So this is a shift.

[00:53:35] Well, it's only – I think that there are certain things that we are definitely adjusting that have only appeared in the last year. A lot of stuff post-pandemic appeared that I think maybe got too aggressive. So I don't disagree. I could like reboot back – by the way, I wrote a whole book on this. So, like, what I – it's not like I have to rescind anything in my book. You know, the things that are in my book are what I actually believe, and those are the things that we're leading with. And you can read that book.

[00:54:04] It's called Trailblazer, and it's about how to bring these core values into your business. And that is extremely important to me that I run a business. Look, like one of the best decisions I made was 26 years ago when we started Salesforce. We put 1% of our equity, 1% of our profit and time into a foundation. It's worked out really well. We've done 10 million hours of volunteerism, given away a billion dollars, run 50,000 nonprofits and NGOs for free in our service.

[00:54:31] And we also started to pledge 1%.org. We've recruited 20,000 other companies to do this with us all over the world. So, they've given away billions of dollars to their local communities. That's important to me. These are the things that are important. And, by the way, I'm not just building products. I am also building a company and a culture. And that's a critical part of what we're doing every single day.

[00:54:57] And you worked in a company who I have a lot of respect for the CEO, Rupert Murdoch, you know, at News Corp. And, Tom, you were there. And it's a different set of values. And, like, Rupert and Robert Thompson, who I know incredibly well, brilliant CEO, you know, I have a lot of respect for different cultures and different businesses. I work with them, especially as you go around the world, different countries.

[00:55:23] But I want to have my company be a certain kind of company built on trust, customer success, innovation, equality, and sustainability. Those things are important to me. You're not going to have a hard time ripping those out of my heart. That's who I am. Right. I think that's probably as good a place as any for us to end here for my sake and for Eric, because I'm getting increasingly urgent messages from your people, Mark. Oh, okay. Well, I hope you got what you wanted. Yeah, we had a great time.

[00:55:49] And I just wanted to say, you know, journalists, you know, can be a little bit like, oh, I don't have any point of view. I think, you know, at Newcomer, we have much more of a perspective. And certainly the 111 program is an inspiration. And, you know, we do – I do applaud you for what you've done in instilling of value. And it's certainly, I think, in a world where company – you know, people want meaning so much. And, I mean, you run sort of a software company.

[00:56:18] And the fact that you can bring meaning as part of that work I think is really valuable and something a lot of people in tech need to figure out. But even if I want you to be more of a warrior on some of these Trump issues, I'm happy about how you're building your company. And we appreciate you coming on the podcast. All right. Oh, no. I'm very grateful to be with you. And, look, I'm happy to do it any time. This doesn't have to be a one and done. All right. Thanks, Mark. Thanks so much. See you. Bye, guys.