Frontier models now discover and chain vulnerabilities on their own, following a low severity bug across services until it becomes a full account takeover. Traditional scanners bury those paths under thousands of unranked findings, and attackers reason across your architecture far faster than any triage queue can. In this session, Rubrik's own security engineers show what becomes possible when you turn that same offensive AI capability inward, against your own code, before anyone else gets there, and rank what you fix by what an attacker would actually exploit. Something new is coming. Be in the room.

And as AI reshapes both how we build and how we’re attacked, resilience becomes even more fundamental. If an agent can touch the data that runs your business, that data must be recoverable. Mike will take that on directly in the AI track.