When the Insider Threat Gets an Agent

Every year, Pavel Livshiz sits across from CISOs and asks them what keeps them up at night. And every year, the answer is the same. Not a sophisticated nation-state attack. Not a zero-day. The biggest risk, consistently, is someone already inside the building: an employee who accesses something they shouldn't, sells their credentials, or walks out the door with permissions they were never asked to return.

‍

That problem isn't going away. It's just about to get a lot more complicated.

‍

"That used to be a human," Pavel said in a recent conversation with Dr. Oren Fuerst on Channel 10. "Soon it'll be an agent."

‍

The reason is straightforward: agents are being created across organizations faster than anyone can track them. "Today anyone can start creating them. You want to know which agents exist in the organization, what permissions they have - because you want to know what they're allowed and not allowed to do, without stopping the work of the organization."

‍

On the data infrastructure side, Pavel sees a similar pattern playing out. "We believe the data stack is going to go through the same transformation that the cybersecurity stack went through 10-15 years ago. It'll just do it faster, because everything moves faster."

‍

And on the longer horizon, he's watching post-quantum cryptography closely, framing it the way he frames regulation generally. "Regulation is sometimes not a bad educational tool. We remember GDPR; it created a wave of companies." PQC, he thinks, is next.

‍

The full conversation, in Hebrew, is below.

‍

‍