Microsoft Enhances Copilot with AI-Driven Deep Research Tools

26.03.2025

Microsoft is introducing two new AI-driven deep research tools, Researcher and Analyst, within Microsoft 365 Copilot. These tools are designed to enhance users’ ability to conduct in-depth research and advanced data analysis using sophisticated reasoning models. With the rise of AI-powered research assistants across various platforms like OpenAI’s ChatGPT, Google’s Gemini, and xAI’s Grok, Microsoft aims to differentiate itself by integrating these tools with both web-based data and internal work-related information. This approach allows users to access a broader range of insights while maintaining relevance to their specific business needs.

Researcher leverages OpenAI’s deep research model, the same foundation behind ChatGPT’s deep research capabilities, but enhances it with advanced orchestration and deep search functionalities. Microsoft claims that Researcher can handle complex analytical tasks, such as developing go-to-market strategies and generating comprehensive quarterly reports. What sets it apart is its ability to pull information from various data sources, including third-party platforms like Confluence, ServiceNow, and Salesforce. By connecting with these external tools, Researcher enables users to access a richer, more comprehensive dataset for decision-making, making it a powerful asset for business professionals and researchers alike.

Analyst, on the other hand, is built on OpenAI’s o3-mini reasoning model and is optimized for handling complex data analysis. It progresses through problems iteratively, refining its responses step by step to improve accuracy. Analyst is also capable of executing Python scripts, allowing it to process intricate data queries and generate more precise insights. This ability to expose and justify its reasoning makes Analyst particularly valuable for users who require transparency in their AI-generated results. By breaking down large datasets and systematically working through them, Analyst aims to reduce errors and improve reliability in data-driven decision-making.

While these AI models offer significant advantages, they are not without challenges. Like other reasoning-based AI systems, Researcher and Analyst may still generate incorrect citations, draw inaccurate conclusions, or rely on unreliable sources. Microsoft acknowledges these limitations and is working to refine the tools to minimize errors. Ensuring that AI-generated insights remain trustworthy is a critical factor, as misinformation or data inaccuracies could have significant consequences for businesses relying on these tools. Despite these concerns, Microsoft believes that integrating AI-powered deep research tools with enterprise data will create a more efficient and intelligent research experience for users.

To provide early access to these new capabilities, Microsoft is launching the Frontier program, which will allow Microsoft 365 Copilot customers to be the first to experience Researcher and Analyst. Those enrolled in the program will gain priority access to these features starting in April. Moving forward, Frontier participants will also receive early access to future experimental Copilot features, positioning them at the forefront of Microsoft’s AI advancements.