OpenAI launches Data agent in ChatGPT Work for natural-language enterprise analytics

OpenAI launches Data agent in ChatGPT Work for natural-language enterprise analytics

OpenAI launched a Data agent in ChatGPT Work that connects to enterprise data warehouses, investigates metric changes in natural language, and builds interactive dashboards.

OpenAI launched a dedicated Data agent in ChatGPT Work on September 10, 2026, allowing business teams to investigate performance metrics, query corporate databases, and build interactive dashboards through natural language. Delivered as a workspace plugin, the tool queries approved enterprise data warehouses while preserving existing corporate permissions. The launch expands ChatGPT Work beyond text collaboration into direct data exploration and reporting. 123

What launched

SignalConfirmed detailAction window
Direct warehouse connectivityConnects natively to Amazon Redshift, Snowflake, Databricks, Google BigQuery, ClickHouse, MongoDB, and Datadog, alongside document context from Google Drive and SharePoint. 23Audit active database connectors in workspace settings and establish service-account boundaries.
Semantic layer integrationBinds metric calculations and data relationships directly to authoritative models including Databricks Genie Ontology, dbt, GitHub repositories, and Snowflake Horizon. 2Register standardized corporate metrics and calculation logic before rolling out broad team access.
Interactive dashboard generationGenerates visual dashboards with filtering, chart customization, and layout controls inside the chat session, with export paths to Tableau, Power BI, Sigma, ThoughtSpot, Omni, and Oracle BI. 23Evaluate interactive dashboard outputs against existing scheduled business intelligence reports.
ChatGPT Sites distributionUsers can publish dashboards to workspace teammates as live ChatGPT Sites artifacts, with optional automated data refresh schedules. 23Review data distribution policies, as published sites store static data snapshots accessible to designated viewers.
Enterprise permission inheritanceQueries enforce role-based access control, table-level restrictions, row filters, and column masks configured on the connected data provider account. 23Confirm user account entitlements and verify that sensitive columns remain masked in test runs.

Analytics workflow and governance constraints

The Data agent guides users through exploratory data analysis without requiring custom SQL queries. A user prompts the agent to identify drivers behind anomalous metric changes, and the system examines underlying dimensional splits, isolates contributing factors, and constructs visualization layouts. Once a team verifies an analysis, users can instruct the agent to dispatch summary findings to Slack channels, generate executive briefings, or suggest operational follow-ups. 23
Enterprise administration follows a two-tier configuration model. Workspace administrators enable the Data plugin under workspace settings and manage installation policies for specific roles. Individual users must authenticate through their existing corporate data warehouse and business intelligence credentials. The plugin cannot bypass provider-level access controls: if an employee lacks access to a specific schema or table in Snowflake or BigQuery, the Data agent cannot retrieve those records. 3
Deployment carries definite limitations. The Data agent is available only to organizations on ChatGPT Work, Business, Enterprise, and Education plans. Publishing reports to ChatGPT Sites creates a snapshot of the underlying dataset, requiring administrators to govern site sharing carefully to prevent unauthorized internal data exposure. 23

Why it matters

The Data agent brings natural-language query interfaces into corporate warehouses while grounding responses in curated semantic layers. Rather than generating ungrounded SQL against raw tables, the agent ties metric definitions to verified sources such as dbt and Databricks Genie, reducing calculation drift across departments. By combining data ingestion, chart composition, and cross-platform export into a single conversational session, OpenAI targets the backlog of ad-hoc analytics requests that typically occupy business intelligence teams. Enterprise IT organizations should verify semantic definitions and audit workspace plugin permissions before permitting widespread exploratory use.

This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.

Related content