Agent Verifiability: The Missing Layer in AI GRC

The podcast explores the governance, risk, and compliance (GRC) challenges presented by the increasing use of AI agents within enterprises. It highlights the need for automated frameworks to manage and monitor these agents, drawing an analogy to the management of vast networks. Radhika Venkatraman emphasizes the importance of verifying agent identity and actions, while Sriram introduces a categorization of agents into single-player, multiplayer, and zero-player types, each requiring different levels of control and governance. The discussion also covers the necessity of continuous monitoring and the potential use of blockchain for creating immutable audit trails. The panel also touches on the balance between deterministic code and non-deterministic human input in AI, as well as industries that are most likely to adopt AI agents.