Okay, so we've talked about the blueprint for the secure adjective enterprise. We've talked about Okta for AI agents and now we're going to ground it in one of those real concrete conversations with an amazing company and amazing leader at an amazing Okta customer and this is a conversation with John Rose from Dell. John, come on up here. So let's start. Let me put Dell in the AI context. How do you think about the industry? Dell's place in the industry, how the industry is changing. Where does Dell fit into there?
Yeah, we're fairly central because you cannot do AI with the infrastructure of the past. Our business is building out the IT infrastructure of the world: compute, storage, networking, and all things associated with it. That's been good for business. We're doing very well. We are finding ourselves in an entirely new world. However, just building infrastructure for someone else would be difficult because you wouldn't know what to build. So a couple of years ago we made the decision to be customer zero. We would be the early aggressive adopter of this technology. I took over as the chief AI officer about two years ago, having been the CTO for quite a long time.
Was there another person or is that a new position? There was a person there for a very brief period of time that attempted to build consensus and didn't work.
We decided to go top-down. We tried bottom-up, didn't work. We went top-down when I took over. That fundamentally launched our journey: if we're going to build for this new space, we sure as heck better understand what the space is as a first-hand participant in it. We can spend a long time talking about our journey, but the punch line is after two years of a very disciplined approach, we just had our financial reports. In those two years, something happened that hadn't happened in 41 years: our revenue grew dramatically, first year about 10 billion, second year about 20 billion, and at the same time our costs went down. We have never seen that happen. When you redesign for the AI era, people, process, technology, you decouple revenue and cost because the unit of work is no longer just a human being. It's an augmented human being or an agent. That gives you tremendous leverage to grow your business and improve your cost structure. Doing that is non-trivial. We leaned in to be bleeding edge but also pragmatic. One interesting statistic: for all that progress, we never used a single agent yet. We built our first autonomous agents almost two years ago, but the ones in production are minimal. However, we are extremely excited about the shift to autonomy. We believe AI can only be applied to a process. With first-generation tools, we could only go after simple processes tied to human work. In the agentic world, we can go after complex, composite, and autonomous processes, opening up a gigantic surface area that will define productivity for the next decade.
When you think about the Agility technology stack, what are the key technologies? There's process, change management, strategic direction, and technology enablement. You mentioned a bunch of stuff was done without even Agility. What are the key parts of the Agility stack?
Yeah, you and I talked a while back. At the end of last year in October, after working with these technologies, we wrote it down. We have an internal set of Agility standards. They're very specific. They included things like we decided all MCP servers would be centralized because MCP is too risky to use without putting it in a safe space. We made decisions about having two kinds of Agility platforms: inner ring, which we wholly own and operate, and outer ring, which exist within our partners and SaaS services, but they are part of one topology under one architecture. The single biggest decision we made is that every agent that does work on behalf of Dell would carry a digital identity issued and operated by us. I don't care if the agent is in a third party or a SaaS provider. I want control. I authorize it to exist. Because if I control identity, I can do things like a kill switch. I can make an agent go away if it behaves badly, even if it isn't running on my infrastructure. That stack has formed: a modern AI factory underneath, a knowledge layer with knowledge graphs and RAG systems, and on top a control plane that includes identity and access management and authorization. That stack is real and published. We started to provide it as a pseudo RFI to all of our partners. We think the industry ought to take blueprints and use them.
Yeah, it sounds simple. I'm an identity guy. You guys early on made the decision that agents have to have first-class identities. We saw the survey I mentioned earlier that only 22% of companies are managing agents as real identity things. It's good to hear that you're seeing success, but it's a subtle thing that a lot of people don't get. They try to treat them as more like software and scripts and service accounts. It's a very profound thing that we can make happen together.
So, when you talk about agents, you guys decided the primary platform and the use cases, a lot of them are internal. How do you think about agents operating internally for your enterprise versus agents outside your enterprise accessing your systems and the internal external trade-off there?
Yeah, I would say our first thoughts about agents over a year ago, we had this belief that we could build one platform and run all the agents on it. By the way, we have a bias. We really protect our intellectual property and our data. I am very hesitant to let black boxes of magic touch my data. So I like doing things that I can control, whether in my data centers or in systems that I helped design the architecture for a provider. But we quickly realized that agents track work, and work can exist in many places. Some I fully control, some I don't. So we pivoted to a two concentric ring model: platforms we fully control and platforms we don't control the platform but we can control the agents on the platform. That's where identity and access management became the universal substrate. Interestingly, we don't have full consensus in the industry about what even an agent is.
Wait, no, wait, stop, John.
