What happens when your AI agents become the security risk? In order to prevent that from happening, organizations need proper governance of them. That’s why the company introduced Agent Cloud, which is powered by something called SAGE, or Semantic AI Governance Engine.
In this interview we recorded during RSAC 2026 Conference, we go into why the agentic AI revolution is creating unprecedented security challenges that traditional identity management simply cannot address. Together with Dev Rishi, Rubrik’s GM for AI, we discuss how agents are already operating in enterprise environments with creativity and autonomy that breaks conventional security models.
What you’ll learn in this video
Rubrik believes AI-powered governance is the only scalable approach to securing agentic systems. Its SAGE framework addresses challenges that seemed unsolvable just months ago:
- Why existing identity management systems fail when agents interact laterally with multiple credentials
- How do you deal with semantic complexity?
- How SAGE learns from human feedback to prevent alert fatigue while maintaining security
- The technical architecture behind deploying small language models for real-time policy enforcement
- Why the shift from rules to semantic understanding represents a fundamental change in security thinking
- Real examples of how agents creatively circumvent controls designed to restrict them
- The three essential capabilities every enterprise needs to manage agentic operations