Microsoft Adds Advanced AI Research Tools to Copilot Suite
Microsoft is set to expand the capabilities of its Microsoft 365 Copilot AI assistant with the launch of two new tools designed to enhance deep research and data analysis, as reported by Tech Crunch. The tools, Researcher and Analyst, bring advanced reasoning models and data integration features to the enterprise suite, positioning Microsoft to compete with similar offerings from OpenAI, Google, and xAI.
Elevating Research with AI
The Researcher tool combines OpenAI’s advanced research model, the same technology used in ChatGPT’s deep research function, with Microsoft’s orchestration and search technologies. The result is a research agent capable of producing complex reports, like client-ready quarterly summaries or go-to-market strategies. Microsoft says the tool’s strength lies in its ability to synthesize information from various sources, both internal and external.
Meanwhile, the Analyst is designed for more technical and iterative data work. Built on OpenAI’s o3-mini reasoning model, Analyst is tailored to perform detailed, step-by-step data analysis. It uses Python to handle sophisticated data queries and can reveal its internal logic for transparency, which may appeal to IT, data science, and compliance teams.
Access to Internal and External Data
Microsoft’s tools differ from other deep research agents like Google’s Gemini or xAI’s Grok because of their dual access to internal work data and public web sources. Researcher can connect to enterprise platforms such as Confluence, ServiceNow, and Salesforce through third-party data connectors. This allows users to gather insights not only from the internet but also from proprietary business systems.
Also read: Itron Teams Up with Microsoft to Revolutionize Utility Operations with Generative AI
A Frontier for Experimental Features
Microsoft plans to roll out Researcher and Analyst to select users through a new Frontier program to provide early access to experimental features within Microsoft 365 Copilot. Starting in April, organizations enrolled in the program will be the first to explore the new tools.
Despite the promising capabilities, Microsoft acknowledges that even these advanced tools are not immune to hallucinations, a common issue in AI where models fabricate information, misattribute sources, or base insights on unreliable data. The company notes that while tools like Researcher and Analyst are built with fact-checking and self-correction mechanisms, occasional errors may still occur.
