OpenAI launches Data agent in ChatGPT Work to turn enterprise data into dashboards

OpenAI is pushing ChatGPT Work deeper into business analytics with a new Data agent that lets employees ask questions of approved company data and turn the results into dashboards. The product, announced on September 10, is designed for teams that need answers from operational, financial, sales or customer data but do not want every request to wait on a specialist report or a manually written query.

The Data agent connects ChatGPT Work to data sources that companies already use, including Amazon Redshift, Datadog, Google BigQuery, ClickHouse, Databricks, MongoDB and Snowflake. It can also use files and documents from Google Drive and SharePoint. The important part is not just access to raw tables. OpenAI says the agent can draw on business definitions, metric logic, custom calculations and data relationships from semantic layers and trusted systems such as Databricks Genie Ontology, dbt, GitHub, Snowflake Horizon and existing BI dashboards. That context is meant to reduce one of the biggest problems in enterprise analytics: two teams asking the same question and getting different answers because they use different definitions.

The workflow sounds closer to an analyst than a search box. A user can ask why sales slowed, where spending is rising or which support issues threaten renewals, then follow up in the same conversation. The agent can inspect evidence, refine the analysis and turn findings into an interactive dashboard with visualizations. OpenAI says the dashboards can be edited, shared and refreshed, and that organizations can apply their own brand guidelines to the output. The agent also works with business intelligence tools such as Omni, Oracle BI, Power BI, Sigma, Tableau and ThoughtSpot.

Governance will determine whether the product is trusted. OpenAI says enterprise administrators decide which data connections are available and which roles can use them. Queries are supposed to enforce the user's existing permissions, including table, row and column restrictions. That matters because the value of conversational analytics depends on access to sensitive information, but a single permission mistake could expose customer, employee or financial data to the wrong person.

The Data agent also hints at how enterprise AI is moving beyond chat. OpenAI says users can ask ChatGPT Work to recommend next steps, identify who should be involved and share findings through connected tools such as Slack or email when approved. In practice, that positions ChatGPT as a layer that can move from question to analysis to action, while still relying on company systems for identity, permissions and source data.

For the analytics market, the announcement raises pressure on BI vendors and data platforms. Natural-language querying has been promised for years, but many attempts struggled because they lacked trusted business context or could not produce reviewable evidence. OpenAI's approach is to combine language models with semantic layers, governed connections and dashboard generation. If it works reliably, more employees may treat analytics as a conversation they can drive themselves. If it fails on governance or definitions, companies will keep human analysts in the loop for the most important decisions.