Microsoft’s Copilot is even better at research thanks to new features that combine the power of both OpenAI’s ChatGPT and Anthropic’s Claude. In a blog post announcing the availability of Copilot Cowork, Microsoft debuted the Critique feature used in Microsoft 356 Copilot’s Researcher tool. Unlike standard Copilot, Researcher is designed to tackle more complex tasks in multiple steps.
Now, Researcher is even better at that with its Critique feature that uses GPT responses. GPT responses were refined by Claude. “This architecture creates a powerful feedback loop that delivers high-quality results across factual accuracy, breadth of analysis, and presentation,” Microsoft said in a blog post, adding that Researcher’s processes are similar to those found in “academic and professional research settings.” Microsoft claims that the upgrade scores slightly higher on the Deep Research accuracy, completeness, and objectivity benchmarks (compared to the latest Perplexity Deep Research model). Anthropic has a research feature that allows you to use multiple Claude agents to provide comprehensive responses to more complex requests.
If you want a little more autonomy in your research, Microsoft has also added the Model Council feature, which is an alternative option to Researcher. With Model Council, you get answers from both Anthropic and OpenAI in parallel, and a report showing where your models match and disagree. Both capabilities are currently available in Microsoft 365 Copilot’s Frontier program, which serves as an early access space to the company’s AI innovations.
