Microsoft is actively testing ways to bring OpenClaw-style autonomous agents into its Copilot AI assistant, The Verge AI reports. The goal: make Microsoft 365 Copilot “run autonomously around the clock” while handling tasks for users without constant supervision.
Omar Shahine, Microsoft’s corporate vice president, confirmed to The Information that the company is “exploring the potential of technologies like OpenClaw in an enterprise context.” That’s a significant admission from a company that typically builds its AI tools behind closed walls.
What’s OpenClaw and Why Does It Matter?
OpenClaw is an open-source platform that lets users create AI agents running locally on their devices. It surged in popularity earlier this year, but brought serious security concerns along with it. Microsoft believes it can build “safer” versions of the technology, according to sources cited by The Information.
The always-on Copilot could reportedly:
- Monitor your Outlook inbox and calendar automatically
- Serve up a daily list of suggested tasks
- Run role-specific agents for marketing, sales, and accounting
- Limit agent permissions by siloing them from other business areas
That last point is worth highlighting. Instead of giving one AI agent the keys to everything, Microsoft is exploring role-based agents with restricted access. A marketing agent wouldn’t touch your accounting data. A sales agent stays in its lane. This approach directly addresses the security nightmare that made OpenClaw controversial in the first place.
The Bigger Picture
This move signals a shift in how Microsoft thinks about Copilot. Until now, Copilot has been largely reactive. You ask it something, it responds. An always-on agent flips that dynamic entirely. Copilot would proactively identify work, suggest actions, and potentially execute tasks while you’re focused on something else.
Microsoft reportedly plans to showcase some of these features at its Build conference starting June 2nd. That gives the company about seven weeks to polish what it wants to show developers and enterprise customers.
The timing matters. Competition in the enterprise AI space is intensifying. Last year, Anthropic launched Claude integrations inside Microsoft 365 services and brought its Claude Cowork tool to Copilot for handling “long-running, multi-step tasks.” Google has been pushing its own Gemini agents into Workspace. OpenAI is aggressively courting enterprise customers.
As The Verge AI notes, bringing OpenClaw-like capabilities into Copilot could help Microsoft reclaim customers it lost to rival services.
What This Means for Enterprise Users
If Microsoft pulls this off, the workflow changes are substantial. Instead of manually triaging your inbox every morning, an AI agent could have your priorities sorted before you sit down. Instead of bouncing between apps to prep for meetings, Copilot could assemble briefing docs overnight.
The security question remains the biggest hurdle. Enterprise IT teams are already cautious about AI access to sensitive data. Autonomous agents running 24/7 with access to email, calendars, and business documents will face intense scrutiny before any large organization gives them the green light.
Microsoft’s bet is that enterprise-grade guardrails and role-based permissions will make the difference. Build 2026 should reveal whether that bet has substance or is still mostly ambition.
More details are available in the original report from The Verge AI.