MongoDB Atlas plugs into Microsoft 365 Copilot: live operational records answer questions without a data migration

Live MongoDB Atlas operational data can now be queried conversationally inside Microsoft 365 Copilot, across Copilot Chat, Researcher in Copilot, Copilot Cowork, and Word, Excel and Teams, with additional experiences planned to follow. Previously, the information needed to answer questions such as which incidents are open or which orders are delayed typically lived in operational applications built on MongoDB Atlas, while the questions were asked in everyday productivity tools, leaving a gap between the two. The connector is generally available, and users can use natural-language prompts to retrieve and filter customers, transactions, products, orders, content or operational events, get counts and key performance indicators such as delayed orders or high-value customers, and summarize and rank data.

The connector brings Atlas into Copilot through MongoDB's managed remote Model Context Protocol (MCP) server, so when a user asks a relevant question, Copilot queries Atlas in real time instead of relying on a static copy of the data. The MCP server acts as the standard query channel between Copilot and Atlas, letting business users without MongoDB expertise access and analyze live operational data through an AI assistant. Using the connector requires no migration of operational data, no re-platforming of applications and no rebuilding of integrations; an administrator approves and enables it in the Microsoft 365 admin center, and after users sign in, their existing Atlas permissions carry through to each query.

Atlas remains the system of record, and the connector queries Atlas directly rather than indexing a copy of Atlas data into Microsoft 365, so the information returned reflects the current state of the underlying application data. This federated approach connects the productivity layer to operational data while keeping that data in the environment where it is managed, without adding a data-movement project or another application to maintain. Each query runs as the signed-in user and honors permissions already defined in Atlas, the connector is designed for read-only interaction, administrators can audit connector activity in Microsoft Purview, and there is no separate search index.

The demonstrated scope is read-only querying, filtering, counting, summarizing and ranking of live Atlas data through the federated connector across Microsoft 365 Copilot experiences, with queries honoring existing Atlas permissions; it applies to organizations already running MongoDB Atlas on Azure and using Microsoft 365 Copilot.

The connector is designed for read-only interaction and does not describe writing or modifying Atlas data; additional Copilot experiences are described as planned without a listed scope; the release gives no quantitative data on query latency, usage or business outcomes.