Microsoft Unleashes Copilot Cowork: AI Models Now Check Each Other's Work with 7 Point Leap

Microsoft rolls out Copilot Cowork, a feature that enables AI models to handle multi-step tasks and check each other's work, with a 7 point performance leap.

Key Takeaways

Microsoft's Copilot Cowork feature is now more widely available as part of the Frontier program.
The feature enables AI models to handle multi-step tasks, access files, and perform calendar planning.
Microsoft's Researcher tool now includes a Critique function where one AI model reviews another's work.

What is Copilot Cowork

Imagine having a team of AI assistants working together to get tasks done more efficiently.

Copilot Cowork is a feature that builds on Claude Cowork, allowing the system to handle complex tasks using various tools.
It can access and output files, and even help with calendar planning and daily briefings.
This feature is now available as part of Microsoft's Frontier program, which is designed to bring the latest AI innovations to users.

How the Researcher Tool Works

The Researcher tool is a powerful AI assistant that just got a major upgrade.

The tool now includes a Critique function, where one AI model writes a draft and another reviews it, pulling from both Anthropic and OpenAI models.
This feature enables AI models to check each other's work, ensuring that the output is accurate and reliable.
The result is a 7 point performance leap, outperforming Perplexity with Claude Opus 4.6.

Introducing the Model Council Feature

Sometimes, it's helpful to get a second opinion, or even a third or fourth.

The Model Council feature allows users to compare answers from different AI models side by side, to see where they agree or differ.
This feature provides a more nuanced understanding of the AI's decision-making process and can help build trust in the technology.
It's all part of Microsoft's effort to make AI more transparent and accessible to everyone.