Microsoft has introduced Critique and Council, two advanced multi-model capabilities within the Microsoft 365 Copilot Researcher tool. Critique is constructed as a dual-model system: one model generates the research draft, planning, sourcing, and synthesizing information, while a second model reviews and refines the output, emphasizing source reliability, completeness, and strict evidence grounding. This architecture, based on evaluations across the DRACO benchmark, delivers measurable gains over traditional single-model solutions, surpassing Perplexity's Claude Opus 4.6 by 13.88% in overall research quality.
Introducing Critique, a new multi-model deep research system in M365 Copilot.
— Satya Nadella (@satyanadella) March 30, 2026
You can use multiple models together to generate optimal responses and reports. pic.twitter.com/m4RlQmCKzs
Council enables simultaneous report generation from both Anthropic and OpenAI models, followed by a side-by-side comparison and a summary from a judge model that highlights consensus, divergence, and unique insights.
These features are currently available to users through the Microsoft 365 Copilot Frontier program, targeting professionals and enterprises who require robust, in-depth research capabilities. The system is accessible to organizations enrolled in the program, focusing initially on users in regions where Microsoft 365 Copilot is supported.
New in M365 Copilot: Council.
— Satya Nadella (@satyanadella) March 30, 2026
You can run multiple models on the same prompt at the same time, so you can see where they align and diverge, and understand what each adds. pic.twitter.com/2p7O14OLFp
Microsoft, the company behind these updates, has positioned Copilot as a central element of its productivity suite, integrating leading AI models from multiple vendors. By leveraging both in-house and third-party AI expertise, Microsoft continues to pursue improvements in workplace automation and research reliability, aiming to set a new standard for enterprise-grade research assistants.
