Hello, and welcome back to Inc.'s 1 Smart Business Story. Microsoft is betting that the key to more trustworthy AI research isn’t about better prompts. New upgrades to its 365 Copilot platform pair one AI model that generates research with another that critiques it in real time. The result, Microsoft claims, is more accurate, balanced, and useful reports that outperform rival tools on key benchmarks. The company is also experimenting with having multiple AI models debate the same task before producing a final answer. Alongside this, Microsoft is quietly rolling out a new Copilot feature that can work directly with your company’s files. Does this mean more businesses will adopt AI assistants at work. 

In this article you’ll learn: 

BY BEN SHERRY, STAFF REPORTER

If you’re looking for more trustworthy AI-driven research, check out these new Microsoft features.

Microsoft has announced some major upgrades to its 365 Copilot platform, including to the latest version of its Researcher agent, and the soft launch of its Copilot Cowork feature, built in partnership with AI titan Anthropic.

In a March 30 blog post, Microsoft said that it added new features to Copilot Researcher, the company’s version of the report-generating deep research agents offered by OpenAI and Anthropic. These agents typically work by developing a plan for how to conduct research into a given topic, then creating multiple sub-agents to check hundreds of sources simultaneously. Copilot’s updated Researcher agent takes a slightly different approach.

 Microsoft said that a new “Critique” mode directs the Researcher agent to use two separate models: one to generate the research report, and another to act as an expert reviewer. “By giving evaluation as much emphasis as generation,” Microsoft wrote, “this architecture creates a powerful feedback loop that delivers higher-quality results across factual accuracy, analytical breadth, and presentation.” 

Critique mode will be automatically enabled when using Copilot’s Researcher agent, but users can also choose to have a single OpenAI or Anthropic model handle the entire research process.

 With this new Critique mode, Microsoft said, Copilot Researcher now tops a Perplexity-developed benchmark that judges deep research models’ accuracy, completeness, and objectivity. Additional information on how Microsoft evaluated the new mode can be found here. 

In addition, Microsoft announced a new “Council” mode for the Researcher agent that simultaneously uses an Anthropic model and an OpenAI model to complete the same task, side-by-side. “Once both reports are generated,” the company wrote, “a dedicated judge model evaluates the reports to create a distilled summary of key findings and highlights where the models meaningfully agree or diverge.” 

Microsoft also announced said that Copilot Cowork, a feature based on Anthropic’s Claude Cowork tool. is now available for early adopters through the company’s Frontier program. The program enables users to access tools that are still in development and not quite ready for launch. 

Anthropic released Claude Cowork in January. The feature enables Claude to access and manipulate files and folders on your computer, and to use applications like PowerPoint and Excel. You can think of Cowork as a digital assistant that can use the local information stored on your computer. 

Microsoft announced Copilot Cowork in early March, describing it as a feature that brings the agentic skills of Claude Cowork tool to the 365 Copilot platform. The company said at the time that Copilot Cowork is capable of searching through and manipulating all of your work data, including “emails, meetings, messages, files, and data.”

Unlike Claude Cowork, which takes over your computer to search through your files, Copilot Cowork runs entirely in the cloud, and can only access the information that your company keeps in digital locations like Sharepoint and Outlook. To get early access to Copilot Cowork, business users will need to verify that their organizations have opted in to the Frontier program.