Gallery

Everyone's asking how to govern AI agents. Microsoft's answer with Agent 365 is simple: give the agent a real identity.

When you build on an agent identity blueprint, the Agent 365 CLI provisions a Microsoft Entra Agent ID. Not a service account, not an app registration you have to babysit. A first-class identity in your tenant, governed like any human user.

And that one identity unlocks three things automatically:

Entra ID governance. Conditional Access, identity protection, and access reviews all apply to the agent.

Purview. Every data interaction respects your sensitivity labels, DLP, and retention policies. No extra code.

Defender. Agent behavior gets monitored for anomalies, with the same alerts and response workflows as any user.

This covers user-delegated agents and autonomous ones, the moment the blueprint is created. Nothing to configure separately.

If agent sprawl is on your radar, this is the pattern that keeps it governable.

Getting started docs: https://learn.microsoft.com/en-us/microsoft-agent-365/developer/get-started

#Agent365 #MicrosoftEntra #AIGovernance