Hajra Naz
Microsoft is introducing a new AI-powered “deep research” tool in Microsoft 365 Copilot. This tool is aimed at enhancing research capabilities within its chatbot app.
Recently, several AI-powered research tools have emerged, including OpenAI’s ChatGPT, Google’s Gemini, and xAI’s Grok. These tools rely on advanced reasoning AI models, which can analyze problems and verify information, key skills for conducting in-depth research.
Microsoft’s version includes two components: Researcher and Analyst.
-
Researcher combines OpenAI’s deep research model with advanced orchestration and deep search features. It can assist with tasks like developing go-to-market strategies and creating quarterly reports.
-
Analyst is intended for sophisticated data analysis and is based on OpenAI’s o3-mini reasoning model. It iterates through problems step by step, refines its responses, and can even run Python to handle complex data queries while allowing users to review its process.
What sets Microsoft’s deep research tools apart is their ability to access work data and the web. Researchers can pull information from third-party data connectors like Confluence, ServiceNow, and Salesforce, offering more comprehensive insights.
However, challenges remain. These tools are susceptible to errors, much like other AI models; they may pull inaccurate information, misread data, or quote the wrong sources.
Microsoft is establishing the Frontier program for Microsoft 365 Copilot users to provide them with early access to these features. Starting in April, Frontier participants will be the first to test Researcher and Analyst, along with other experimental Copilot features.
