At the center of the update is a feature called “Critique,” which enables Copilot to use multiple artificial intelligence models within a single workflow. Instead of relying on one system, the tool now combines outputs from OpenAI’s GPT models and Anthropic’s Claude.
Under this setup, GPT generates an initial response, while Claude reviews the output for accuracy and quality before it is delivered to the user. The company said it plans to evolve the feature further, allowing both models to review each other’s responses in a two-way process.
According to Microsoft, the multi-model integration is intended to address one of the biggest challenges in generative AI—so-called hallucinations, where systems produce incorrect or misleading information. By introducing cross-checking between models, the company aims to deliver more reliable and higher-quality results.
In addition, Microsoft announced a new feature called “Council,” which will allow users to compare responses from different AI models side by side, offering greater transparency and control over outputs.
The updates come as the tech giant pushes broader adoption of its AI tools, including the Copilot Cowork agent, which is being rolled out to users in its Frontier program for early access to experimental features. The move reflects growing demand for more autonomous AI systems capable of handling complex workflows.
Microsoft’s latest enhancements also highlight intensifying competition in the AI space, where rivals such as Google with its Gemini models and other emerging AI agents are rapidly advancing their offerings.
Despite the product momentum, Microsoft’s stock rose modestly by about 1% on Monday but remains on track for its weakest quarterly performance since the 2008 financial crisis, reflecting broader investor caution around the pace of returns from AI investments.