Chaudhry says AI-era threats cannot be fended off with legacy security and require a fundamentally different approach.

But the one point I want to make is when AI comes in, you can't keep on doing old type of security to fend off these new threats. You need to look at a different approach.

Chaudhry argues that endless patching is insufficient and organizations must rethink defense.

Everyone is focused on patching and patching and patching. No one is done patching ever. So while patching is important, I think you need to look at things differently.

Chaudhry pitches hiding applications behind Zscaler's cloud exchange so attackers cannot discover or attack them.

The new approach is if your applications can be hidden behind some kind of cloud and exchange that we built, bad guys can't even find you. If they can't find you, they can't attack you.

Chaudhry says zero trust is needed to stop lateral spread once attackers breach one connected office.

when all these offices are connected to each other, the bad guys get into one office. For that, you need to bring zero trust. If we do these two approaches in addition to patching, we'll be in a good shape.

Chaudhry says enterprises, not consumers, must provide security guardrails, especially against AI agents.

I don't expect consumers to be very tech savvy and cyber savvy. I expect banks or all these retail sites to be more savvy. The better security guard rails are being put by enterprises so that consumers can be protected.