Episode 2: Agentic AI Governance

Agentic AI governance is the set of controls that keep autonomous agents accountable as they operate: who owns them, what data they can reach, what actions they may take, and what happens when they drift. It's not a document filed before launch — it's a live function that runs alongside the agent, enforcing policy at runtime. A risk officer walks in and discovers 47 AI agents running in production. 12 are touching sensitive customer data. Three have no owner at all. That's the gap MLOps was never built to close — it keeps models healthy by tracking accuracy, versioning and deployment, but it was never designed to keep an autonomous actor accountable. Agents don't just predict; they take actions with real consequences, they compose and cascade by calling tools and spawning other agents, and they run continuously instead of waiting for a retraining window. In this video, Michael Rahm, Director of Product Marketing at Collibra, breaks down a simple, three-move framework for governing agents at runtime: 1️⃣ Structure — register every agent at deploy time, straight from your CI/CD pipeline, so nothing ships without an owner 2️⃣ Operate — put a live AI Trust Score on every system and enforce access policies as code, at the data layer, where agents actually reach 3️⃣ Oversee — give leadership a real-time portfolio view of every agent, scored and owned, with a kill switch that pauses any agent instantly The question isn't whether one of your agents will go rogue. It's whether you'll find out from your monitoring — or from your customers. 📖 Get the full runtime control-plane framework, including how agentic AI governance differs from traditional AI governance and MLOps, and how to start governing the agents already running in your environment: https://www.collibra.com/blog/agentic-ai-governance-a-control-plane-framework-for-governing-autonomous-ai-agents-at-runtime #AgenticAI #AIGovernance #AIAgents #Collibra #RuntimeGovernance #ResponsibleAI #AICommandCenter #DataGovernance