Microsoft Agent 365 is the control plane for AI agents in an enterprise, giving IT a registry to observe, govern and secure them.
Overview #
Microsoft Agent 365 #
Microsoft Agent 365 is the control plane for the agents an organisation is accumulating, whether or not IT approved them. Microsoft Agent 365 provides a single registry in the Microsoft 365 admin center showing agent adoption, activity and health, then adds the governance around it: lifecycle management and access control through Microsoft Entra, information protection and data loss prevention through Purview, and threat detection through Defender. The premise is that agents acting on a user's behalf need the same identity and compliance treatment people get.
Production credibility: Microsoft states that Agent 365 became generally available for the Commercial segment on a per-user basis on 1 May 2026, and documents it on Microsoft Learn, where the overview was last updated on 23 October 2026. Microsoft notes that it works best with Microsoft E5 as a prerequisite and that at least one user must hold a qualifying Agent 365 licence to enable it. Partner agents can be deployed from the Microsoft 365 admin center at launch. Microsoft publishes plans and pricing on its own product page; third-party figures circulating for per-user cost are not confirmed by Microsoft's documentation.
Key Features
- Centralised agent registry in the Microsoft 365 admin center
- Agent map and role-specific oversight for security and business leaders
- Lifecycle management: install, publish, block, delete or reassign agents
- Entra-based access control for agents acting on a user's behalf
- Purview for information protection, DLP and risk safeguards
- Defender for threat detection and real-time protection
Ideal Use Case
Enterprises where agents have arrived faster than policy — copilots, partner agents, and whatever individual teams have wired up — and where security needs one place to see what exists, who owns it and what it can reach. It fits organisations already standardised on Entra, Purview and Defender, since the value is in reusing those controls rather than adding a parallel stack.
How Microsoft Agent 365 differentiates
Zenity, WitnessAI and Lakera approach agent risk as a security product you buy alongside your platform. Agent 365 is the platform's own answer, which makes it the default for Microsoft-standardised enterprises and largely irrelevant elsewhere. Infisical comes at the same problem from the credential side. The practical question for a buyer is coverage: Agent 365 governs agents it can see and register, so its usefulness depends on how much of your agent estate lives inside Microsoft's perimeter.
FAQ
Q: What is Microsoft Agent 365? A: It is Microsoft's control plane for AI agents in an organisation, providing a central registry to observe agent activity and the governance and security controls to manage them through Entra, Purview and Defender.
Q: When did Agent 365 become generally available? A: Microsoft states it became generally available for the Commercial segment on a per-user basis on 1 May 2026.
Q: How is Agent 365 licensed? A: Per user. Microsoft says it works best with Microsoft E5 as a prerequisite and that at least one user must hold a qualifying Agent 365 licence to enable it. Plans and pricing are published on Microsoft's product page.
Q: What can administrators do from the registry? A: Install, publish, block, unblock, delete or reassign ownership of agents, and view adoption, activity and agent health across the organisation.
Q: Does Agent 365 cover agents that are not Microsoft's? A: Microsoft says partner agents can be deployed from the Microsoft 365 admin center at launch, and governance extends to agents the platform can register.
tl;dr
Microsoft Agent 365 is the enterprise control plane for AI agents: a central registry in the Microsoft 365 admin center plus governance and security through Entra, Purview and Defender. Generally available since 1 May 2026, licensed per user, working best with E5.