Summary
OpenAI walks through its new Data Agent inside ChatGPT Work, showing how the tool turns everyday business questions into trusted, data-backed plans. The demo follows a product launch scenario: a manager asks about adoption trends from their phone, and the data agent pulls from the company’s warehouse, Slack conversations, and Google Drive documents to surface an analysis, complete with visible SQL queries and metric definitions the data team can audit.
The video highlights how the data agent builds shared dashboards so product, customer success, and data teams work from the same numbers, and how it can be asked to set up alerts — for example, flagging a week-over-week drop in creator adoption — with configurable thresholds and notification rules. It also shows how ChatGPT can turn recommendations into an actionable plan with assigned owners.
A key piece of the demo is the underlying ‘context layer,’ a shared set of business definitions (what counts as a site view versus a site creation, reporting periods, comparison rules) that the data team builds once so future questions across the organization return consistent answers. For businesses evaluating how AI agents can support data teams and reduce ad hoc reporting requests, this offers a concrete look at OpenAI’s enterprise data tooling in action.
📺 Source: OpenAI · Published September 18, 2026
🏷️ Format: Tutorial Demo
