MongoDB Atlas is now a connected data source in OpenAI's Data Plugin, powered by the MongoDB Atlas Managed MCP Server. Analysts and business teams can point the Data agent in ChatGPT Work at the data in their Atlas clusters and ask what changed in the moment, why it changed, and what to look at next, then turn the answer into a report or real-time dashboard without leaving the conversation.
MongoDB Atlas is the data platform applications run on, and where orders, sessions, and events land in real time. Connecting Atlas to the Data agent means an analyst can question the real-time system of record instead of waiting on last night's ETL job to finish. When a metric moves at 2 p.m., the answer doesn't have to wait for tomorrow's refresh.
Turning questions into actions
Here's what this looks like in practice. With the Data agent for ChatGPT Work, you can:
Monitor how the checkout update you shipped at 2 p.m. is impacting conversion minutes and hours later. You can also see where users are dropping off, and decide whether to roll the update back.
Find orders stuck in payment pending over the last hour, see what they have in common, and hand the list of affected customers to support.
Mid-sale, compare this hour to the same hour of the last sale, spot the region that's lagging, and move promotion and inventory before the sale ends.
See what a customer did in the app in the ten minutes before they call support.
Turn these insights into live data charts and dashboards for teams to view in real time.
MongoDB Atlas supports these use cases as a fully managed, real-time operational data layer that runs across cloud providers and operates as the live system of record for their mission-critical applications. Consequently, the data most worth interrogating tends to live in Atlas first. Teams that already trust MongoDB Atlas to run applications in production can now question their data directly.
Real-time operational data helps teams understand what's happening in their business while there's still time to act. With MongoDB Atlas connected to the Data agent in ChatGPT Work, business users can ask questions about that data, investigate changes, and turn the results into reports, dashboards, and recommended next steps–all with their organization's existing access controls in place.
”Meanwhile, end-user access respects the controls already on the cluster. Access is granted through an OAuth sign-in to MongoDB Atlas with no connection strings or credentials to self-manage. Organization owners choose whether to enable AI client access (it's off by default), can restrict it to read-only, and every query runs with the permissions of the user who connected.
Next Steps
Read our press release to learn more, or get started by installing both the MongoDB plugin and the Data agent plugin.
Not registered yet? Sign up for a free Atlas account and start building today.