
Microsoft moved Agent 365 to general availability on May 1, 2026, positioning it as a unified control plane to discover, govern, and secure AI agents across enterprise environments.
Standalone pricing is $15 per user per month, with the platform also bundled into the new Microsoft 365 E7 suite at $99 per user per month.
The launch makes Microsoft the first hyperscaler to ship a dedicated governance layer for agentic AI. The framing is deliberate: Agent 365 is not a tool for building agents. It is a security layer for registering, governing, and protecting every agent.
What it actually does at GA
Agent 365 is built around three pillars: observe, govern, and secure. At GA, organizations can:
Defender context mapping, which charts relationships between agents, devices, MCP servers, identities, and cloud resources, arrives in public preview in June 2026.
The governance gap it doesn't close
Agent 365 is strongest inside the Microsoft ecosystem. Organizations running agents on AWS, GCP, or non-Microsoft frameworks get registry import, but not the same depth of policy enforcement available for first-party agents.
Shadow AI outside managed Windows endpoints also remains a gap. An employee accessing Claude, ChatGPT, or any browser-based AI tool on an unmanaged device sits outside Agent 365's visibility entirely.
The MIT AI Agent Index found that 40% of deployed agents have zero safety monitoring, no logging, no access controls, no behavioral constraints. Agent 365 addresses the managed fleet.
CloudEagle.ai addresses that surface, providing cross-vendor AI tool discovery through browser activity, finance signals, and identity provider data, covering both sanctioned deployments and the shadow AI usage.
Let the headlines come straight to you with Access Granted — a monthly edition of Okta announcements, expert perspectives, analysis, and more.
