OpenAI has introduced a new Data agent in Chat GPT Work to help employees across businesses answer data questions without needing technical expertise. The agent connects to approved company data sources like Amazon Redshift, Google BigQuery, Snowflake, and Databricks, and can also pull files from Google Drive and SharePoint. It uses organizational context such as business terms, metric definitions, and data relationships from trusted sources like dbt, Snowflake Horizon, and BI dashboards to ensure accurate interpretation.

Users can ask follow-up questions, investigate results, and turn analysis into interactive dashboards with built-in visualizations that can be shared, edited, and refreshed. The agent respects existing permissions, enforcing table, row, and column-level restrictions set by administrators. It integrates with tools like Power BI, Tableau, Slack, and email to share findings and execute approved actions.

OpenAI reports that nearly all of its product team and over two-thirds of its GTM organization now use the Data agent internally. Early adopters in the Alpha program include NTT Data, Thermo Fisher, and Service Piston, who use it to analyze sales, detect reporting errors, and prioritize opportunities. To get started, administrators can install the Data agent via Workspace settings > Plugins in Chat GPT Work, connect data sources, and begin conversations with @Data using prompts like diagnosing metric changes or designing KPI frameworks.