Shadown AI: How to discovery unauthorized AI agents in M365

In this video, I explain in a practical way what Shadow AI is and why this topic is becoming increasingly critical for Security, Compliance, IT, and Microsoft 365 administration teams. We are not only talking about users accessing AI tools through the browser. We are talking about local AI agents that can interact with files, commands, applications, code, browsers, and corporate data — often without approval, governance, or visibility from the organization. Throughout the video, I explain the new Shadow AI experience in the Microsoft 365 Admin Center, how this capability helps detect and block AI agents such as OpenClaw, what the prerequisites are, the required licensing, RBAC permissions, the dependency on Microsoft Intune, and the key precautions before applying policies in a production environment. If you work with Microsoft 365, Intune, Defender, Entra ID, Purview, SOC, Compliance, or AI governance, this video is for you. In this video, we cover: What Shadow AI is The difference between using AI and using AI agents What OpenClaw is How Microsoft is bringing visibility to unmanaged agents Prerequisites and licensing Required RBAC permissions Detection and blocking through the Microsoft 365 Admin Center and Intune Key precautions before applying blocking policies How to prepare an AI governance strategy for the organization The main question is simple: Does your company know which AI agents are running on its endpoints?