Microsoft Enhances AI Assistant with Multi-Model Technology for Better Accuracy

Microsoft announced significant enhancements to its artificial intelligence assistant on Monday, introducing technology that enables users to harness several AI models working in tandem during a single task, marking the company’s continued effort to strengthen its AI capabilities and increase user adoption.

The tech giant introduced a new capability called “Critique,” which allows Copilot’s Researcher agent to draw from both OpenAI’s GPT and Anthropic’s Claude models when generating each response, moving away from dependence on just one model.

Under this system, GPT creates the initial response while Claude examines the output for precision and quality before users see the final result, according to Microsoft. The company plans to eventually make this process work both ways, enabling GPT to also review responses generated by Claude.

“Having various different models from different vendors in Copilot is highly attractive – but we’re taking this to the next level, where customers actually get the benefits of the models working together,” Nicole Herskowitz, corporate vice president of Microsoft 365 and Copilot, said in an interview with Reuters.

According to Herskowitz, this multi-model strategy will accelerate user workflows, reduce AI hallucinations where systems create incorrect information, and deliver more dependable results, ultimately enhancing both productivity and quality for customers.

The company is simultaneously introducing “model Council,” which enables users to view and compare responses from various AI models displayed alongside each other. These improvements accompany Microsoft’s broader rollout of its new Copilot Cowork agentic AI tool to participants in its “Frontier” program, which gives customers early access to cutting-edge AI capabilities.

Microsoft had previously introduced Copilot Cowork in testing phase earlier this month, a tool inspired by Anthropic’s popular Claude Cowork product, as the company seeks to capitalize on rising interest in autonomous AI agents.

The software giant continues working to enhance its Copilot assistant to increase user adoption while facing fierce competition from competitors like Google’s Gemini and independent agents such as Claude Cowork.