New agent connects corporate sources, creates analyzes and dashboards, and preserves access controls. See how to validate accuracy and governance.

Direct answer

On September 10, 2026, OpenAI introduced the Data agent in ChatGPT Work. The proposal is to connect data and context approved by the company, transform questions into analyzes and interactive dashboards and trigger flows with existing access controls. The conversational interface reduces barriers, but does not automatically make data trustworthy: catalog, lineage, permissions, defining metrics and reviewing decisions are still necessary.

Simple question can hide complex rule

Revenue, margin and active customer need shared definitions for the answer to be comparable.

Access must accompany the source

The agent must not extend permissions; identity and filters by row, column, or workspace must remain in effect.

Citation improves verifiability

Any relevant conclusion must point to the query, period, source, and transformation used.

Dashboard does not close the analysis

Visualizations speed exploration, but causality, exceptions, and quality require expert review.

Nexus Reading

Start with recurring questions, create expected response sets, and measure accuracy, net time, and corrected decisions.

FAQ

What does the Data agent do?

Connects approved business sources and turns questions into analyses, dashboards and actions.

Does it replace data governance?

No. Catalog, quality, access and lineage remain the organization's responsibilities.

How to validate before scaling?

Compare responses to approved queries, record sources, and test permissions and ambiguous cases.

Essential guides to delve deeper into the decision

Primary sources and references

This editorial analysis was produced by Nexus from the official sources below, consulted on October 5, 2026. The text is original and interprets practical implications for companies.

Date reported by the main source: September 10, 2026.