Microsoft is testing AI bots inspired by the OpenClaw platform to enable 365 Copilot to function as an autonomous agent. The company wants 365 Copilot to become an "always-on" agent that completes tasks on behalf of users, according to a report from The Information.
Microsoft Vice President Omar Shahine confirmed the company is "exploring the potential of technologies like OpenClaw in an enterprise context."
OpenClaw is an open-source platform that allows users to create AI-powered agents that run locally on a user’s device. While the platform gained popularity earlier this year, it also introduced several security risks.
Microsoft plans to implement "safer" versions of these tools to mitigate those risks, sources told The Information. The company aims to maintain the benefits of autonomous execution while preventing unauthorized access to enterprise data.
Role-specific automation
Microsoft is also exploring OpenClaw-like agents tailored to certain roles, such as marketing, sales, and accounting, to "limit the permissions the agent needs," siloing them from other parts of a business, according to The Information.
The proposed autonomous Copilot could monitor Outlook inboxes and calendars to provide daily task suggestions. These agents would proactively manage schedules and surface actionable items to users without requiring manual prompts.
Restricting access to specific datasets ensures that a marketing agent, for example, cannot access sensitive accounting or legal files.
Microsoft aims to showcase these features during its Build conference, which begins June 2. The showcase will highlight how autonomous agents can handle complex, multi-step workflows.
Anthropic recently launched integrations with its Claude AI chatbot inside Microsoft 365 services. Anthropic also introduced its Claude Cowork tool to Copilot to help complete "long-running, multi-step tasks."
Microsoft’s development of OpenClaw-like capabilities represents an effort to retain customers who have migrated to rival AI services.
