Microsoft takes Agent 365 out of preview as shadow AI becomes an enterprise threat

Microsoft has made Agent 365 generally available, positioning it as a unified control plane for discovering, governing, and securing AI agents across Microsoft and third-party environments.

The update highlights a growing need to manage locally installed and unmanaged agents, which can contribute to shadow AI and software sprawl inside enterprises.

Microsoft says the platform is designed to help defend against new agent-era risks, including unsecured backend connections, cross-prompt injection, and data access controls that are not agent-aware.

The launch arrives as companies adopt autonomous AI across cloud services and employee devices, increasing the chance of leaks, misuse, and unauthorized actions without stronger governance.