Most teams sit on data they cannot easily question, so Alistair opened OpenAI's mini-series with a tour of the Data Agent in ChatGPT.
A plugin is software that adds outside tools and data to ChatGPT, and Alistair said this one grew out of internal use. He said what started for OpenAI's data team spread fast to marketing, product, engineering, sales and ops. Installation starts in the plugins marketplace, where you search for data and install it. Once on, you can ask a question and get an answer with a visualization. You can also build and share dashboards with Sites and business intelligence tools, or tell the story in an interactive report.
It meets files where they live.
Alistair said the plugin works with files, APIs and custom MCPs, where MCPs are open connectors that let the model pull from outside systems. Skills ship with it for checking data quality, defining and tracking KPIs, product and business analysis, and sizing a market, according to OpenAI. That setup stands apart from narrower helpers. A recent roundup calls OpenAI's Data plugin the connected-data generalist, while Excel and Sheets add-ins stay in the grid and Snowflake handles plain-English SQL. You can also ask ChatGPT to propose next steps, update docs with new data and share findings through Slack, Teams or email.
Install it now and connect your own sources, then watch the next installment on linking favorite data tools. This first part is only the start, Alistair said, so expect more connectors and skills.




