Microsoft has introduced new capabilities for its AI assistant Copilot that will allow users to simultaneously use multiple artificial intelligence models within a single workflow.
Microsoft has introduced new capabilities for its AI assistant Copilot that will allow users to simultaneously use multiple artificial intelligence models within a single workflow.
With the help of a new feature called Critique, the Copilot Researcher AI agent will now be able to draw on results from both OpenAI GPT and Anthropic Claude models for each answer, instead of relying on just one model, Reuters writes .
According to Microsoft, while GPT generates a response, Claude will check the result for accuracy and quality before showing it to the user. In the future, the company plans to make this process two-way, allowing the GPT model to act as a reviewer for Claude's drafts.
“Having a variety of models from different developers in Copilot is great in itself. But we’re taking it to the next level, where customers get real value from having these models work together,” said Nicole Herskowitz, corporate vice president, Microsoft 365 and Copilot.
“A multi-model approach will help speed up user workflows, keep AI hallucinations under control, and provide more reliable results, which will increase customer productivity and quality,” Herskowitz added.
Microsoft is also launching Model Council, a feature that will allow users to compare responses from different AI models side-by-side. These updates come alongside the expansion of Copilot Cowork, a new agent-based AI tool that is now available to a wider range of members of the Frontier program, which gives customers early access to the company’s latest AI developments.
