OpenAI adds ChatGPT Data plugin for work analytics and BI

The new Data plugin is less about raw capability and more about distribution and stickiness. By baking a generalized BI and analytics layer into ChatGPT Work and Codex, OpenAI is trying to capture a workflow that has historically lived in tools like Power BI, Looker or Tableau. If teams can connect their warehouse, CRM and finance systems, then ask Astra-class models to explore trends, explain anomalies and build dashboards on the fly, the model becomes the primary interface to data for many knowledge workers.

Strategically, this moves OpenAI further up the enterprise stack. Instead of just selling tokens or agent platforms, it is now competing in analytics and decision-support, where switching costs are higher and the value per seat is larger. That could pressure both cloud-native BI vendors and competitors like Microsoft and Google, who have been weaving their own copilots into office suites. It also accelerates the normalization of agentic models sitting directly on top of sensitive business data, which will force faster maturation of permissions, audit and lineage controls.

In the broader race to AGI, the Data plugin is a reminder that the path is not only about new model families. It is also about making existing frontier models the default fabric for everyday analytical thinking inside organizations, so that when genuinely general systems arrive, the organizational plumbing is already in place.