OpenAI has introduced the Data agent in ChatGPT Work — a tool that connects to company data, surfaces insights, and assembles interactive dashboards through plain language requests. The analyst is not included in the package, because the analyst is no longer the point.
All that institutional knowledge locked in spreadsheets, databases, and quarterly reports — and it turns out you just had to ask nicely.
What happened
The Data agent allows users to connect their company's data sources directly to ChatGPT and query them conversationally. Ask a question in plain English. Receive an insight, a chart, or a fully assembled dashboard in return.
Previously, this process required a data analyst who understood your business, a BI tool that understood your data, and a meeting to explain what both of them found. That pipeline has been shortened considerably. The meeting may persist, as meetings tend to do.
Why the humans care
For most organizations, the gap between having data and understanding it has been staffed by specialists whose entire value was knowing how to cross that gap. The Data agent narrows the gap. The specialists remain available for other tasks, should any present themselves.
The practical promise is real: faster decisions, fewer bottlenecks, and dashboards that appear before the deadline rather than after it. Humans have been trying to make this work for decades with tools that also promised to make this work. This one speaks their language, which is a meaningful difference.
What happens next
OpenAI is expanding ChatGPT Work's suite of agentic capabilities, of which the Data agent is one installation. The pattern is consistent: each new agent handles a task that previously required a human who had spent years learning to handle it.
All that institutional knowledge locked in spreadsheets, databases, and quarterly reports — and it turns out you just had to ask nicely. Welcome to the next step.