Microsoft introduces a broad range of security updates aimed at helping companies securely manage and protect AI agents as full-fledged components within their IT environment.
The announcements indicate a clear change of course at Microsoft, writes SiliconANGLE. Microsoft no longer views AI agents as ordinary applications, but as a new layer that must be central to the security architecture. Companies are increasingly dealing with autonomous systems that have access to data, applications, and workflows.
Therefore, the tech giant states that companies need centralized control mechanisms for agent management, stronger identity protection, stricter data protection, and security processes that
Centralized control with Agent 365
The most important introduction is Microsoft Agent 365, which will become generally available starting May 1. This platform serves as a control layer for AI agents and bundles technology from Defender, Entra, and Purview. This gives companies an overview of which agents are active, what access they have, and how data is being used. Microsoft is integrating Agent 365 into a new bundle, Microsoft 365 E7, along with Copilot, Entra, and other solutions.
Microsoft is further expanding its suite with, among other things, a Security Dashboard for AI, which centralizes risks surrounding AI usage. New features also detect shadow AI at the network level and show which AI tools are being used within organizations without approval. Additionally, there will be more visibility into software and AI applications via Intune, allowing IT teams to identify and address risks more quickly.
Focus on identity and data
Microsoft is also implementing innovations in the area of identity management. New Entra features are designed to help companies better protect and manage identity through passkeys, backup capabilities, and risk-based access. On the data side, Microsoft is integrating security directly into AI workflows. With Purview, for example, companies can block sensitive information in prompts or monitor how data is used by AI systems.
Notably, Microsoft is also focusing on “agentic defense”: AI that automates security tasks. Security Copilot and new AI agents within Defender and Sentinel are intended to detect and handle threats more quickly.