It fundamentally

changes how we need to think about insider risk, right? Historically, whenever we spoke about an insider, we assumed there was a human being behind the keyboard, someone accessing files, sending emails, interacting with business applications.

Now, Imagine the same employee has five AI agents working under their behalf.

Those agents have access to Salesforce, Google Drive, email, GitHub, and potentially your entire business environment.

And they're not just reading information anymore, they're taking actions, making decisions, executing workflows, sometimes communicating with other agents, and increasingly running autonomously in the cloud.

So suddenly you have this entire new workforce operating inside your enterprise.

And the interesting security challenge, I think, is that all of that activity can happen through perfectly legitimate access, right? The agent has the right credentials, the API calls are authorized, and from the perspective of your existing security infrastructure, everything might look completely normal.

That's a good

What does that statistic tell you, Aviv?

I'll emphasize what I said before.

I think the way we think about the 95 split at above is quite simple.

Most insider risk is not the evil employee.

I mean, that person exists.

He definitely does.

There are malicious insiders.

How is AI reshaping insider risk from your perspective?