As agents gain access to enterprise tools, organizations need consistent ways to govern what those agents can access and enforce those decisions at runtime
Today, we’re announcing a new integration between Microsoft Agent 365 and Azure API Management now in public preview. The integration connects centralized governance in Agent 365 with runtime enforcement for MCP servers and tools through Azure API Management for V2 and AI Gateway tiers.
With this integration, AI administrators can discover and govern gateway managed MCP servers and tools from Agent 365 and application platform teams use Azure API Management, to enforce those access decisions at runtime. Microsoft Defender and Microsoft Purview add additional layers of protection by evaluating tool requests and results for threats and sensitive information risks.
Centralized governance, enforced at the gateway
Agent 365 and Azure API Management play complementary roles. Agent 365 defines governance intent, maintains inventory, and manages access decisions. Azure API Management enforces those decisions inline for traffic to MCP servers and tools.
Together, they connect the enterprise governance experience with the infrastructure platform teams already use to manage access to business systems. Administrators can make supported access decisions from Agent 365, while enforcement happens at the gateway in the customer’s Azure environment.
For agents already using a configured gateway endpoint, that connection remains in place. Customers can extend governance to their existing gateway-managed tools without rearchitecting those agents or moving tools out of their Azure subscriptions.
Discover tools and control what agents can do
The integration brings together four key capabilities:
- Discover existing gateways and MCP servers. After a one-time tenant-level consent, Agent 365 discovers supported Azure API Management resources and the MCP servers they expose, removing manual registration.
- Maintain a centralized inventory. Administrators can find gateway-managed MCP servers in Agent 365 and inspect available metadata, bringing these assets into a shared governance experience.
- Control access at the right level. Block or unblock an entire MCP server. With the Azure API Management AI Gateway tier (preview) they can apply more granular controls by blocking individual tools without disabling the server.
- Use runtime signals for oversight. Tool activity observability on AI Gateway tier supports monitoring, auditing, and incident investigation, connecting access decisions with what happens at runtime.
Add protection before tools run and before results return
Controlling which tools agents can access is one part of the governance model. Organizations also need protection for information flowing through those tools .
The Agent 365 tool threat protection policy (supported in the AI Gateway tier) uses Microsoft Defender and Microsoft Purview capabilities to evaluate tool requests and results for risks such as malicious content, prompt injection, and exposure of sensitive information.
Platform teams can configure evaluation before tool invocation, after invocation, or at both stages. With blocking configured, pre-invocation evaluation can prevent a disallowed call from executing. Post-invocation evaluation can withhold unsafe results before they reach the agent.
Build on open standards and existing investments
This integration connects centralized governance with enforcement where agents access tools, while building on the infrastructure organizations already operate.
MCP provides a common interface for tool access. Azure API Management provides the gateway boundary for applying policy, security, and observability. Agent 365 brings those gateway-managed assets into a centralized governance experience, with Microsoft security capabilities adding protection along the tool-call path.
The goal is not to make every agent use the same runtime. It is to help organizations apply governance to tools reached through supported gateway endpoints, bringing AI administrators, platform teams, and security teams closer together, while preserving the existing architecture.
Get started
API Management v2 tiers support centralized MCP inventory and server-level blocking and unblocking through Agent 365.The AI Gateway tier adds more granular capabilities including individual-tool blocking, tool activity observability, and tool threat protection .
Start with the Agent 365 integration guide, explore tool threat protection with Defender and Purview, and configure your gateway in the AI Gateway portal.