Microsoft is expanding Agent 365 with broader discovery of AI agents and an Action Center for automated governance tasks. The changes give IT teams more ways to find agents and act on risks, rather than simply keeping an eye on them.
Microsoft AI Watch analysis
What happened
Microsoft says Agent 365 will expand discovery to 60 types of local agents, including agents it does not already know about, and add cloud-agent discovery in preview. The company also says more than 50 million agents have been registered in the platform.
The new Action Center is designed to help IT and security teams clean up inactive agents, reassign agents without owners and automatically block agents based on risk. Microsoft also says Agent 365 uses unique agent identities, ownership and purpose information, and access controls, while drawing on tools organisations may already use, including Entra, Defender, Purview and Intune.
Our top picks
- Discovery of unknown agents
Finding agents that have not been registered could give security teams a clearer view of what is running across their estate. - A route from alerts to action
The Action Center is designed to handle tasks such as removing inactive agents and blocking agents flagged as risky. - Identity and ownership for agents
Microsoft says each agent can have a unique ID, an owner and a stated purpose, with access controls tied to its needs. - A broader agent registry
The platform is designed to include Microsoft, customer-built, local and partner-cloud agents, with integrations including Anthropic, Databricks and Amazon Bedrock. - Cost and value management in development
Microsoft says it is working on ways to classify tasks and help organisations understand what agents cost and what value they deliver.
Why it matters
As organisations add agents from different providers and frameworks, knowing which ones exist, who is responsible for them and what they can access becomes a practical management problem. Discovery and automated controls could help teams deal with that sprawl before every agent gets its own corner of the company’s IT estate.
Microsoft says Gartner named it a Leader in the inaugural 2026 Magic Quadrant for AI Agent Management Platforms. That is an analyst assessment, not evidence that the controls work equally well in every organisation. The useful news here is the product direction and the specific capabilities Microsoft says it is adding.
Our read
Agent management is moving from a registry problem towards an operations problem: find the agents, know who owns them, then do something when they are inactive or risky. The Action Center is the most consequential addition because automation could make governance less of a spreadsheet hobby. The test will be whether its recommendations and blocks are useful in real deployments, not merely tidy on a product page.
What to watch
- When discovery of the additional local agents and unknown agents becomes available.
- Which agent types and partner integrations are supported in practice.
- How teams can review or reverse automated actions, particularly risk-based blocks.
- When Microsoft shares more detail on cost and value management.
Discussion spark: Should organisations let agent-management tools automatically block agents they flag as risky, or should a person approve every block?
Sources and evidence
- Microsoft named a Leader in the 2026 Gartner® Magic Quadrant™ for AI Agent Management Platforms – Microsoft (3 November 2026, 17:00 UTC)
not affiliated with or endorsed by Microsoft