Why AI agents are the future of work, with Marc Benioff

Stripe is built for whatever comes next because AI isn't just changing business, it's evolving fraud. Stripe Radar detects fraud with the fast scale of Stripe's 1.9 trillion dollar annual transaction volume. In fact, when a card is used, there's 92% chance the Stripe network has seen it before. That means emerging fraud patterns are detected sooner, blocking bad actors, welcoming good customers. Learn more about fighting fraud with Stripe at stripe dot com slash radar.

On Masters of Scale, iconic leaders reveal how they've beaten the odds.

Asking really strong questions is a superpower.

You want to show up with something radically different and how they've grown companies to incredible heights.

The greatest rewards always come from the greatest risks.

Let's hit the gas. Airbnb, Zillow, Microsoft, Liquid Death, and more. Hear from the founders who've changed the game. It's anything but business as usual. Find Masters of Scale on Apple Podcasts, Spotify, YouTube, or wherever else you get podcasts. The business world is moving faster than ever. And when change hits, we need to learn in real time. On Rapid Response, you'll hear candid conversations with CEOs and leaders making tough calls about AI, human talent, responsibility, and the bottom line.

How they navigate uncertainty, pressure, and high-stakes moments. I'm Bob Safian, former editor-in-chief of Fast Company, and I'll be your host. As each episode breaks down what you need to know right now, you can find Rapid Response wherever you get your podcasts.

I just wrote our new business plan for fiscal year 26 and I do it myself. I've done it for 25 years. I love doing it. And I answer these questions. What do I want? What is important to me? My methods, what is preventing me from having it? My obstacles, my measurements, my KPIs. And as I'm writing the business plan for the year, you know, sometimes I'm working with a colleague as well, and we're doing it in partnership.

Well, for the last three years, I'm also working with an AI.

Marc Benioff is co-founder and CEO of one of the largest software companies in the world, Salesforce. And now they're focused on AI agents that can be deployed in all kinds of workplaces and organizations. These AI agents not only act on our behalf, but also collaborate with us, maybe even help us write our fiscal business plans.

And I then will ask the AI, Hey, what do you think of my business plan? Is it competitive? How does it compare to what my competitors are doing? What should I be doing differently? What is your suggestion? So across all business functions, AI is part of what we are doing to make us more productive, more augmented, and have better customer results and ultimately better KPIs for our business.

On this episode, I'm talking to Mark about Salesforce's new agentic layer called Agentforce, how AI will reshape the future of work, and how agentic AI could revolutionize our governments. I'm Rana El Kaliouby, and this is Pioneers of AI, a podcast taking you behind the scenes of the AI revolution. Mark, welcome to Pioneers of AI.

Oh, I'm so thrilled to be here with you.

Yeah, thank you for joining us. We met, uh, a few years ago at Dreamforce, and so it's great to reconnect. Thanks for being here.

It's great to be back with you.

Salesforce has been using AI in their work for years. But this past fall, at their annual Dreamforce conference, they launched Agentforce, a suite of autonomous AI agents. Think of them as 24/7 customer service reps that can resolve most issues and know when to bring a human agent into the mix. I wanted to begin my conversation with Mark, talking about the scope of Agentforce and what it can actually do.

Well, I'll just give you, you know, I'm the CEO of Salesforce. We'll do about $41 billion this year in revenue, and we have about 75,000 employees. And, you know, we have about 9,000 support agents at Salesforce. You probably know that. And we also have a site called help dot salesforce dot com. And in the last ninety days, if you went to that site, you could work with a agent or you could work with a human.

It's your choice. And our agents completed about 380,000 conversations with our customers in the last ninety days. They did that at 84% accuracy and resolution rate. So there was only a few customers who had a few questions that they had to hit the button and boom, they were over with the human. And so that was very exciting, and that's to me what an agent is. It's a way to help a company, you know, become more productive, augment their employees, grow their revenues using this technology, where really it's about humans and AI and agents working together.

So I want to double click on this idea of of collaboration. My thesis around AI is human-centric AI, so AI should really be about augmenting and amplifying human abilities, and that's how we unlock human potential. How does this vision of agentic AI and Agentforce specifically fit into this? And in particular, like, what's the framework of these AI agents? Are they assistive? are they autonomous? How do you think about all of that?

Well, there's really about three things that are going on here for us. One is, look, we're just a lot more productive and efficient as a company. So while we have 9,000 support agents, when we finish this fiscal year, we'll probably have less than that. Because we need less because our agents are able to do so much more. So one thing, it's about productivity and efficiency, and then two, those human agents are way more augmented. So when those customers who are kind of hitting the upper bounds of AI have to get pivoted over to a human, you know, our service cloud screen comes up and it's a seamless integration between the two.

And they can see all of the conversations that you've had. Now, that means that human is a lot more augmented and has more capability and is able to know more about you as a customer. So think about it like this. We have our applications, those are things like sales, service, field service, marketing, commerce, our platform, Slack, Tableau, all as a common core set of applications, deeply integrated with our data cloud, which is where all the data is, you know, for those applications.

And that data cloud is also federated, which means connected to all the other data sources in your company. And then there's an agentic layer on that. And so it's the agents, the data, and the apps together as one deeply integrated platform, and that deeply unified platform, that's one code base, that's one piece of code. And by the way, this wasn't really possible even a year ago. So we launched this product in, um, we announced it in September, we launched it in October, we started using it ourself in November.

Customers like Remarkable and Equinox have already deployed it. And now we've also signed some amazing customer stories who are doing some incredible work. Uh, one of them is Lennar, the nation's largest home builder. And, um, you know, they are able to service their customers twenty four seven, sell them new products like mortgages and insurance. They're able to provide all kinds of new functionality to help lower their costs, but increase their revenues. This is really the promise of these AI agents, deeply integrated with the data and apps as well.

So there are plenty of other software companies that are releasing or building their own AI agents. It's becoming kind of a pretty crowded space, and you've actually been quite critical about some of them, including Microsoft's Copilot. What makes Agentforce different or unique?

Well, that it's real, you know, you we've heard about Copilot for a few years from Microsoft, but it's really disappointed so many customers. Look, we have to be the best example. We have to be customer zero. That's very important to me as the CEO of Salesforce, that I'm really using our products and that I am created, you know, one of the largest software companies in the world. That is based on our technology, that is, we grew our business using our platform, our customer success platform.

And that is, I think, a good point of evidence.

AI agents are already changing the way Mark and his team work at Salesforce. But how will agentic AI revolutionize work as a whole? We'll get to that after a short break. Stay with us. The companies winning right now aren't just building great AI. They're building the infrastructure to actually run a business on top of it, which is where Stripe comes in. Stripe is built for whatever comes next, with planetary scale financial infrastructure to help you power any business model.

As a matter of fact, 86% of the Forbes AI 50 run on Stripe, along with industry leaders like Salesforce and Pepsi. Last year, Stripe users processed $1.9 1 trillion combined. See how Stripe can power what's next for your business at stripe dot com. Let's talk about the implications of agentic AI, specifically on the future of work. And how do you see agentic AI improve work, both from the perspective of an employee or a professional like each of us, um, as well as the employer perspective.

Well, we are going to be augmented in incredible new ways. I mean, one of our products is Tableau, is something I use to help manage the company, to understand what's going on, to really get into the data, to move from data to decisions. And that's a product a lot of us have grown up with, actually. You know, it's a product that's been around for a couple of decades. It's awesome. It does incredible analytics and business intelligence.

But we rewrote it, and we wrote it, we wrote it over the last, uh, two years, and we rewrote it. So that it was not only a great app for humans, but that it was deeply integrated into our data cloud and again, deeply integrated in our agentic layer. That Trinity I'm going to keep coming back to because we found that that is the key to making our product so successful in the agentic era. So now we have a new version of Tableau, Tableau Next, this incredible new platform that is built on our data cloud and our agentic layer.

This idea that Tableau, a very well trusted business intelligence brand, now has an augmented agent capability. This is really incredible. And when I'm using it, I'm working with the agent to have a more intelligent capability in running my own business.

Yeah, and it probably also lowers the barrier to like as an interface to Tableau and similar tools, right? Because I imagine you can. Can you prompt? Can you just like have a conversation with Tableau asking it about different data insights and graphs and whatnot? Is that the idea?

Yes, you can still have all the power and functionality and capability of Tableau, but you're also then deeply vested into the data cloud. And yes, you have the agentic layer, and that's really important for what we call the data fam. The data fam is the Tableau community that loves Tableau, that's been with it even before we acquired the company. So when we rewrote the product, we knew we had to kind of preserve the kind of incredible, you know, capabilities of Tableau and what people love about it.

But we have to position it correctly for the future as well. And I think we got that one really right.

Yeah, it sounds exciting. So I'm seeing a lot of new business models where the company is running less like a SaaS company and more like an AI staffing organization or an AI staffing agency. It's this idea where you can, like, hire an AI agent to complete or automate a task. That is going to have, like, major implications on, again, the future of work. Some jobs are just going to become obsolete. And in fact, the World Economic Forum came out with a survey recently that indicates that 41% of companies worldwide plan to reduce their workforce by twenty thirty because of AI.

How do you think about and respond to these concerns?

Well, I think that we are in a new world. There's no question. Everybody understands what's going on. And I think we look at the World Economic Forum. They just ran their Davos conference in January. And they use Salesforce to manage all their information and keep track of who their attendees are and what the sessions are. But this year, they used Agentforce, and Agentforce was running side by side with the attendees to help them have a better conference experience.

That I thought was very much representative of what is going to be the future. Yes, you not only have people around the conference helping you get from point A to point B, asking the information booth, Hey, how do I do this? But now you also have an agent to help you get to where you need to get to.

But I want to go back to your example with human agents and Agentforce servicing your customers. You actually said that probably means that you will need a lot less human agents in the future, right? So what happens to all these people who are no longer needed in these jobs? Like, how how is Salesforce thinking about this?

Well, I can just tell you for my own, you know, company, which is I'm hiring still very aggressively right now. I'm expanding my customer organization by ten to 20% this year, and I'm still focused on growing many functions. But there are some functions that I'm not growing, like in engineering. I've achieved more productivity, so I'm holding the headcount flat. And in the example of customer support, yes, I'm reducing because I don't need as many customer support agents.

So I have the ability to do a lot of different things. I can rebalance, I can shift. I'm trying to achieve what is the most competitive company that I can create, and I'm using technology to do that. So that is very much a tool that I have in my arsenal to be able to make this a success. And that is the right way to think about it. I can't be thinking about I have to. If I run the business exactly like I did twenty five years ago, is that going to make me successful?

I certainly don't have the same code, I don't have the same product strategy, I don't have the same business strategy, I'm not in the same geographies, even. I've changed all of those things. I have what I call a beginner's mind. I have what the Japanese would say, shoshin, a new approach. In the beginner's mind, I have every possibility. In the expert's mind, I have few possibilities. And by taking that approach, I need to take that approach with my employees as well.

What is the right employee balance? What is the right approach with agents? Where should I be investing in with agents? Where should I be investing in with humans? And how do I get humans and agents to work together?

What kind of new jobs do you think AI will create or is creating?

My gosh, I mean, you know, I mean, right, take take it from the tippy top, you know, my partner as a CEO is an AI. And then right there for my customer support agents, their partners are AIs as well, agents, and also that data analyst and the person who's doing finance and running finance with Tableau. Their job is now highly augmented through AI. You know, AI is not a world model yet. A world model, even though it's one D word, one letter difference between word and world, a world model is more multi-sensory and is able to see everything going on in the world.

Today, the AI is still very much word-based, but that word-based model can do an incredible amount, and I can make my company a lot more productive and successful with it.

Do you have, um, kind of this idea of multimodal AI on your roadmap? Do you think about voice agents and, you know, agents that can see and have vision or other perceptual abilities?

Yes, that's always been part of our vision of, you know, that our technology tracks very closely with the best AI in the world. And if you've noticed Salesforce's history with AI, we've been aggressively developing and engineering our own AI solutions for more than a decade with our Einstein platform. Some of the first protein models, which were published in Nature, all came out of Salesforce Research, and a lot of the top, most productive and most accurate models are all coming out of Salesforce Research, like CodeGen and many others.

And I think that as we look at into the future, you are going to see more multimodal or multi-sensory models. And that will come right into the enterprise as well, and we'll be there to help bring those along for our customers as that technology is mature. But I think one thing that we've also recognized is that, and this is one of the reasons why we have to be customer zero. We have to be the first ones to deploy and then show our customers that this is real and can operate at scale.

So that companies like Lennar, companies like Singapore Air, Heathrow Airport, or whoever it is going to be remarkable, Equinox, all the people who are using Agentforce now, that they can have confidence that this is the right thing and the right capability. Because look, we've all seen the movies, we know where we're going. Minority Report, you know, we see the vision of what's possible. But where we are right now is not at that level. So we can do some things, but we cannot do everything.

Yeah, so let's switch gears a bit. We're talking about the future of work, and there's a big shift happening at the federal government level, of course, when it comes to work. We're witnessing the efforts of the Department of Government Efficiency, or Doge, to make the government more efficient. They're reducing spending, and it also has a lot of ripple effects in academia and other institutions. Basically, Elon Musk has taken an approach very similar to what he did when he bought Twitter and how he's thinking about the government.

So I have a two-pronged question for you. One is what is your response to concerns about applying a Silicon Valley-esque business strategy to the U.S. government? And two, Agentforce is about increasing efficiency and cutting costs. So do you see everything that's happening at the government level, an opportunity for Agentforce? Do you kind of see an opportunity to bring your AI agents into the government and replace some of these job losses?

Well, you're right on so many levels, which is that there's no question that there's going to be an opportunity in every organization, whether it's a business or a government, to bring in agents to augment the workforce. And it's a digital labor revolution. This is going to move from what traditional enterprise software has been, which is a multi 100 billion dollar TAM into something that I've seen the analysts like, Ark say, between a three and 12 trillion dollar TAM.

And that means that agents are going to be part of everything we do, including running our government or running our companies. And you are also right that, yes, Doge is kind of trying to take the traditional Silicon Valley startup playbook and apply it to the government. We'll see how that works out. They're even using Slack, which is a product, you know, that we make at Salesforce that's also been augmented with AI. And, you know, we've rebuilt Slack.

You probably see that it has AI built into it. It has agents. You can do recap, you can do summarization, and soon all your agents will be living right inside Slack as well, and you'll be interacting with them in Slack and deploying them and cultivating them. And so Slack is a critical part of the future of work. And so all of these things can make a more productive government and a more successful, more automated government, more efficient government.

And we already do that today with a lot of government agencies, including like the Veterans Administration, which provides, you know, support and healthcare for our veterans on a Salesforce platform. So there's a lot of capabilities that we can do for the government as well, and for governments.

What is your advice to, you know, young professionals who are having to be kind of concerned about, you know, the the their careers and and and their jobs? W What's your advice for them so that they are successful in this AI-driven world?

Well, I think is that they should go all in. We built a platform called Trailhead, it's available at Trailhead dot com, to help educate and to tune people up for this world, to reskill and to upskill their capabilities. And I think, you know, you've got to really get in it and use it. I mentioned that I wrote the new business plan for fiscal year 26 for Salesforce in partnership with an AI and a colleague of mine.

And this year, I brought in a colleague that I hadn't worked with before on business plan, and I wanted to show them how to do this. So I think it's very important that people do get educated, they get in it. If you understand it, I think you have some responsibility to mentor and show others. I think that is why Salesforce has to be customer zero on everything that we are doing, whether it's the new version of our service cloud or Sales Cloud, or Marketing Cloud, or Slack, like we've talked about, or any other thing that we're doing.

So that companies can really see, here's a model of how the future works.

Yeah, you know, I'm part of the Young Presidents Organization. It's an organization of over 30,000 CEOs globally, and we're all grappling with the question of how do we ensure that our organizations are AI-driven and AI-first. And I love what you said, which is basically your model, your role modeling, right? How you want Salesforce to embrace AI by doing it yourself. And I love that.

Well, I think it's really important for all those YPO members, they can use, uh, Salesforce YPO, I think itself runs on Salesforce. And I think that for each and every industry, whether it's financial services or airlines or healthcare, uh, life sciences, uh, manufacturing, and a cr- across the world, we're gonna go through this incredible transformation. And so the best thing to do is to try it out, to see what you can and cannot do.

I've said it before on this show, but the best way to get started with AI is by testing it out, especially in a work setting. The potential for agentic AI is astronomical. We'll get to more of that after a break. So I'm a techno optimist and I truly believe AI can help us solve some of humanity's and the world's biggest challenges. What are you most excited about for AI and specifically agentic AI? Maybe outside of Salesforce.

Well, I I'm excited to pursue innovation. Like I've been writing software since I was 15 years old, and I just love software, and I just think it's amazing how far software has come over the last 45 years that I've been hanging out with software and working with software and shaping it and marching it forward. You know, there's so many exciting things happening now, even with with chips, with data centers. We see these huge 100 billion dollar, multi 100 billion dollar data center projects.

We've been rebuilding Salesforce so that we run well, for example, in Amazon and Alibaba and Google, and take advantage of all these really low prices that they're able to offer for deployment today based on the huge amount that they've invested. And how do we make our customers more successful with this technology while preserving the trust? I think this is what it really gets down to. And we want to make sure that we're about trust and customer success, about innovation, equality, and sustainability at Salesforce.

And that also means that we have to really demonstrate to our customers that they're going to be able to have trusted customer success with their customers.

How do you think we should? Uh, I love this doubling down on trust, but how do you think we can build that into both, um, you know, products at a company level, but also at our governments?

Well, I think, you know, transparency. I think the more transparency you can have, the more the more clarity they can have on exactly what's happening, the more trust you're going to have. That's what we do in Salesforce. That's one reason I love Slack. It's one reason we bought Slack. You know, Slack was kind of moving along, they were kind of hitting some ceilings in innovation themselves. They weren't able to operate as well as an independent company, and we really made the case that they would do better as part of Salesforce, which they have done.

You know, we've grown Slack dramatically since we've been able to acquire the company and now innovate with it, get it ready for this AI future. And I think when you look at a company like Slack and what it can do for a organization, it's about transparency. And you can really get a vision for what everyone else is working on, what they're doing. At Salesforce, all of our executives publish their personal business plans in Slack. I publish mine so everyone can see what I'm doing.

That idea, more transparency, that gives you more trust. And I think a little more of that and a little less, you know, things that are maybe obscured. Or we don't know exactly what's happening with our politicians, or we don't know exactly what's happening, you know, in the world, or what they're thinking about. I don't think that goes very well for us. And I think that that's one reason to be optimistic about what's happening, even in our new administration, is we need a little more, a little more transparency into what's really going on.

All right, final question. I know you are very spiritual and I think you meditate every day. Is that correct?

That's true.

So okay. So one question that I've become really obsessed with, um, and I would love to hear your point of view is, um, in this age of AI, what does it mean to be human?

Well, I think it's what you just answered the question, which is that, you know, humans have tremendous capabilities that AI may never have. You know, AI is not going to suffer, so it will never really know compassion. That that was something a discussion that I had with somebody who was a great teacher to me many years ago. That idea that, you know, we're, we're, we're amazing, but we do suffer a lot as human beings. And because of that suffering, we become much more compassionate and empathetic towards others because we understand what they're going through.

Uh, we should always remember that, you know, technology is not good or bad, it's what we do with it that matters. You know, one of the things I'm most proud about of 25 years at Salesforce is we took 1% of our equity, our profit, our time and put it into a charity. And that's one of the reasons we've been able to deliver more than $1 billion in grants. We run 50,000 nonprofits and NGOs for free in our service, and we, um, have done 10 million hours of volunteerism, giving back and service to the communities that we do business in.

I think that that's incredibly important. And I think that that idea that we have to be more linked to compassion and service, giving back as we move into this new world is can set a set of values that can guide us very clearly and i improve our ability to ultimately be successful with this technology.

I love that. What a wonderful way to end. Not just about driving efficiency, but really building compassion and serving at the end of the day. This is great. Thank you for joining us, Mark. This was awesome.

It's great to be with you. It's so much fun to see you again.

There's a sea of AI agents entering our workforce, and Marc Benioff is putting Salesforce at the very front of that wave. It's exciting. And yes, some jobs may no longer exist as AI continues to ramp up. But I believe, like Mark, that AI has the potential to augment our abilities. Most jobs will evolve with AI, helping us become more productive and focus on the work that matters. What struck me the most from my conversation with Mark is how he's going all in on AI, at both the company level and individually, as the leader of a multi-billion dollar company.

This sets the tone for the rest of the organization. It's a very powerful stance to take, sending the message to his team, his customers, and stakeholders that this is the future that Salesforce is betting on. We're definitely going to continue talking about agentic AI and where it's being used. There's so much here to uncover, and we're just getting started. Next week on Pioneers of AI, we have another awesome conversation for you. We're talking to one of the pioneers revolutionizing the way AI can improve women's healthcare.

Before you go, don't forget to rate Pioneers of AI and leave us a review wherever you're listening. We read all of your reviews, plus it helps others find the show. Thank you. Pioneers of AI is a Wait What original. Our executive producer is Eve Tro. Our producer is Rachel Ishikawa, and our associate producer is Jordan Smart. Our senior talent executive is Stephanie Stern. Mixing and mastering by Ryan Pugh. Original music by Ryan Holliday. And our head of podcasts is Lital Molad.

You can join the conversation on LinkedIn, Instagram, TikTok, YouTube, and X. Just search for at Pioneers of AI. Thanks so much for listening.