On March 30, 2026, Microsoft unveiled a new feature named Critique for the Researcher product, which is part of Microsoft 365 Copilot. This innovative addition enables GPT and Claude to work in tandem on research-related tasks. Specifically, GPT takes charge of generating the preliminary draft, while Claude meticulously reviews the content, ensuring its accuracy, comprehensiveness, and the quality of citations in line with academic norms. Looking ahead, this feature is set to incorporate bidirectional peer review capabilities. In other words, Claude will have the option to craft the initial draft, with GPT subsequently stepping in to perform the review. Moreover, Microsoft has introduced a multi-model "council" mechanism. Under this system, multiple models conduct independent research, and a judge model steps in to compare and evaluate the outcomes, refining both points of consensus and disagreement. This move signifies the transition of AI assistants from single-model operations to multi-agent collaboration, ultimately contributing to improved content quality and enhanced work efficiency.