OpenAI has launched a Data agent for ChatGPT Work, positioning the product as a way for employees to investigate company data and build shareable dashboards through a conversation rather than a query editor.
The announcement is notable because it puts a data agent next to the systems that define and control enterprise data. OpenAI says the agent can connect to approved sources including Amazon Redshift, BigQuery, ClickHouse, Databricks, MongoDB, Snowflake, Google Drive, and SharePoint. It can also use business terms, metric definitions, calculations, and data relationships from semantic layers and connected tools.
The access model is the product boundary
OpenAI says enterprise administrators choose which connections and roles are available. Queries enforce the connected account’s existing table, row, and column restrictions. That is the relevant detail for a data agent: a conversational interface does not replace the permissions model underneath it.
The agent can build interactive dashboards and work with dashboards in tools including Power BI, Tableau, Sigma, ThoughtSpot, Omni, and Oracle BI. OpenAI also says users can review the evidence behind individual findings and continue the investigation in the same conversation.
Those claims describe a broad workflow, not an independent measure of analytical reliability. The company includes customer examples from its Alpha program, but those are reported experiences rather than a benchmark of accuracy or business impact.
Self-service analysis raises the bar for data definitions
Making natural-language analysis available to more employees can reduce the wait for a report. It can also make a weak metric definition travel faster. If a company has inconsistent revenue logic, stale customer mappings, or unclear exclusions, a polished dashboard will not fix that underlying ambiguity.
