Microsoft 365 Copilot Researcher is a deep research agent designed to help you create more reliable, well-sourced reports inside the Microsoft 365 environment. It brings together multiple AI models and Microsoft 365 capabilities so you can:

• Ask complex research questions in natural language.
• Get structured, citation-rich responses.
• Work within Microsoft 365 security, privacy, and compliance boundaries.

Researcher uses Work IQ along with Microsoft 365’s built-in security, privacy, and compliance features, so your research process stays aligned with your organization’s governance standards while you explore and synthesize information.

Critique is a new multi-model capability in Microsoft 365 Copilot Researcher that helps you review and refine AI-generated content before it becomes a final report.

Here’s how it works:

• You start with a response generated by a GPT model.
• Critique then uses a Claude model to evaluate that response for:
  • Accuracy – Are the facts correct and consistent?
  • Completeness – Are key points or perspectives missing?
  • Citation integrity – Are sources properly referenced and aligned with the claims?

The outcome is a research report that is more accurate, complete, and well-attributed, helping teams trust and reuse the content with greater confidence.

Council is a multi-model feature in Microsoft 365 Copilot Researcher that lets you compare how different AI models respond to the same prompt.

With Council, you can:

• Run one prompt and see responses from both a GPT model and a Claude model side by side.
• Quickly spot where the models agree and where they diverge.
• Identify what each model uniquely adds to the analysis or narrative.

This side-by-side view helps you rethink how you choose AI-generated content. Instead of relying on a single model, you can compare perspectives and select the output that best fits your research goals, improving both the quality and transparency of your final deliverables.