Microsoft has confirmed that the first agentic functions will soon arrive in Windows 11, in the first phase through Insider test versions, and the idea is that AI agents will get their own “workspace” in which they can independently perform tasks on your files.
The key innovation is called Agent Workspace. It is a special area in the system where the AI agent works, connected to the agent’s own user account. While you normally use the computer, that agent in the background can, for example, sort documents, prepare files, perform automated tasks or execute scripts that you assign to it, without you constantly clicking everything manually.
Microsoft emphasizes that agent accounts are separate from the main user account. Agent workspace should consume little processor and memory, and in terms of security it functions similarly to a light virtual machine: the agent is isolated, works in parallel and has limited access only to certain folders such as Documents, Downloads, Desktop, Pictures or Videos. The idea is that you can see and control exactly what the AI has access to.
The company claims that security and monitoring are a “priority”, so agent accounts must have clearly defined powers and must be able to be monitored by both users and security software. However, Microsoft itself admits that AI agents increase security risks, because any autonomous program that works alone in the background can become the target of an attack or perform a task incorrectly if someone uses it.
For now, agentic functions will be marked as experimental and only available to volunteers in the Windows Insider program, but the message is clear: Microsoft wants Windows 11 to eventually become a system in which you no longer manage only you, but also a whole team of AI “workers” who automate everyday tasks. How much users and companies will be ready to trust such agents remains to be seen.