Transcript
enterprises in a massive way is agents. Agents will outnumber everything else. It will be interconnected. It will be accessing things everywhere. We're not going to be able to discern the difference between the two, especially with the tools we use today. It's all going to look like one mesh. Think about your firewalls. Is your firewall going to be agentic aware at all? There's no way. Your endpoint security today is not agentic aware. It doesn't know whether you're doing something versus an agent's doing something. This becomes a really, really scary thing because it's all going to mesh together and agents are going to exponentially change our enterprises and our life. Very rapidly. Like I said, a lot of people are sitting in the room and are like, I don't know about that. Who's scared? Who's scared of agents? Who's scared? Who's scared of agents right now? All right. I think that just like I sat in the room and with those CIOs and those CISOs and said, they said that cloud thing's not happening. I'm seeing the same thing on agents. Not AI, but agents. Now look at this like an autonomy slider and think about if you have kids. You let them. You're like, all right. You can drive the house. Okay. You can ride your bike around the neighborhood without me watching you, but you can't leave the neighborhood, right? You put these controls in place, these guardrails, but it's an autonomy slider and it's the same thing with AI. How much do I allow you to do things at the same time trust you to do that and what guardrails I'm going to put in place? Now the beginning here is just autocomplete of code. If you talk to most developers, they'll say I'm getting somewhere between 10 and 30% benefit to programming. I'm autocompleting. I'm looking at the one that provided me. I'm getting some value. I'm moving on. Now, if I ask our marketing team, we're getting 50, 60% benefit of ChatGDP writing articles from a content creation, but they can't go to 95. If they went to 95%, it would look sloppy. It would look like AI wrote it. There'd be a lot of em dashes and emojis and it would look like AI. You're sitting here, so you have to find this balance of how do you learn to use this technology with controls. The best example to me of really 99.99% autonomy is the Waymo car. You're riding in this thing, you're trusting it with your life. Other people are trusting it with their life. The pedestrians are trusting it with their life, but every once in a while, a person sitting in an operation center has to engage and activate and manage this car. There's still a human in the loop to manage the risk, but all the guardrails. This is 15 years of maturity and development in place, but ultimately, if we can do this with cars, how can we not automate most of the things that you do every day in managing data and managing security and managing risk?
