Microsoft is reportedly exploring OpenClaw-style AI features for Copilot that could make the assistant more proactive inside Microsoft 365.
Microsoft is reportedly exploring OpenClaw-style agentic AI capabilities for Copilot as it pushes to make its AI tools more proactive inside the workplace. According to The Indian Express, citing a report from The Information, the company wants Microsoft 365 Copilot to evolve into a system that can operate more autonomously and complete tasks on behalf of users.
The report says Microsoft is considering always-on upgrades that would allow Copilot to scan a user’s Outlook inbox and calendar, then generate a list of suggested tasks each day. This would mark a shift from assistant-style AI that waits for prompts toward software that actively helps organize workflows and surface priorities.
Microsoft is also said to be working on OpenClaw-like agents designed for specific business functions, including marketing, sales, and accounting. These agents appear to be aimed at enterprise customers, with reduced approval requirements and tighter isolation from other parts of an organization.
A key issue around OpenClaw-style systems is security. As The Indian Express notes, OpenClaw’s rise has also raised concerns because such agents may need broad access to user data and systems to function effectively, creating risks around file tampering or accidental deletion. The report says Microsoft believes it can develop a safer enterprise version of these capabilities.
The move comes as Microsoft appears to be refining its Copilot strategy. The Indian Express report notes that the company has recently been scaling back broad Copilot expansion in favor of integrating AI where it is most meaningful, echoing comments from Windows and Devices EVP Pavan Davuluri.
If introduced, these features could strengthen Microsoft’s position in enterprise AI by making Copilot less of a chatbot and more of an active workplace agent. The larger question will be whether Microsoft can balance automation, usefulness, and security well enough for businesses to trust these systems at scale. This is likely to become clearer if the company formally showcases the features at Build in June 2026.
Microsoft is testing OpenClaw-like AI bots for Copilot
Microsoft tests new generation AI
Microsoft tests OpenClaw style AI in 365 Copilot for automation
Anthropic has released Claude Opus 4.7, its most capable publicly available model to date, delivering significant improvements in advanced software engineering, high-resolution vision, and long-horizon agentic tasks.
Meta launches Muse Spark, a closed proprietary AI model with tiered reasoning and 3B+ user reach across WhatsApp, Instagram, and Quest VR.
Key takeaways OpenAI has signed a multi-year, $10 billion agreement with AI chipmaker Cerebras Systems to secure computing infrastructure. The deal will deliver 750 megawatts of computing power throug...
