So what does it look like when they press forward with, let's just call it irresponsible AI? What kinds of concrete problems have people brought to you and said, hey, we broke it, help us fix it?

Hmm.

Oh, this we have absolutely seen.

I would say, keep in mind that urgency, the pace, the network teams already had these pace and urgency problems.

The AI tools that the attackers are getting just increased that urgency and the need to go to deal with these things at scale.

One of the most common things that we're seeing is customers working hard to bring tools to bear that help them answer basic questions.

The side effect of what we're seeing is if you take a bunch of network automation tools in your environment, wrap them with AI, you not only get that capability to ask questions like, why are some of my backups failing every night? Or at what rate are those failures increasing? Or something like that, that's a normal thing to chase.

nodding quietly, so yeah.