Microsoft has launched Agent 365, a comprehensive control plane to observe, govern, and secure AI agents across local, cloud, and SaaS environments, including agents built by Microsoft and ecosystem partners.

The service supports prebuilt agents in Microsoft 365 Copilot and Teams, Copilot Studio or Foundry agents, and partner-built agents, with integrations from Genspark, Zensai, Egnyte, Zendesk, and factory providers like Kasisto, Kore, and n8n.

Agent 365 unifies identity, security, governance, and management for AI agents across enterprises, launched on May 1, 2026, to enable secure and compliant AI deployment at scale.

Security enhancements extend Entra ID to AI agents for secure authentication and add Defender-powered threat detection to monitor anomalous AI behavior and reduce data-breach risks.

Current visibility includes Frontier programme users seeing OpenClaw agents and their devices, with plans to add a registry of local agents to Defender and Intune for shared inventory.

Governance and management tools come with built‑in compliance templates aligned with GDPR and ISO 27001, plus policies for AI decision-making and lifecycle oversight from deployment to decommissioning.

The service addresses rising oversight and security challenges from growing numbers of AI agents by adding discovery and management through Microsoft Defender and Intune to curb shadow AI.

Agent 365 enables secure scaling of AI initiatives across industries like finance and healthcare, with monetization opportunities through subscriptions and third‑party integrations.

Public previews planned for mid-2026 will introduce asset context mapping to link each agent to its device, configurations, identities, and reachable cloud resources to aid security investigations.

Network controls extend Entra to Copilot Studio and local agents, enabling traffic inspection, outbound connection restrictions, malicious prompt-based attack blocks, and protection against internet and SaaS risks.

These network controls also apply to end-user devices, supporting policy-based screening of file movements and blocking risky activity at the network layer.

FAQ highlights that Agent 365 integrates identity, security, governance, and management tools, extends Defender and Entra ID to AI agents, and points to monetization, healthcare and finance use cases, with integration challenges and long‑term industry-standard impact by 2030.