OpenAI Introduces Data Agent in ChatGPT Work to Analyze Company Data

OpenAI on September 10, 2026 introduced a new Data agent in ChatGPT Work, a plugin the company says connects to approved company data sources, investigates what changed in business metrics, and builds interactive dashboards employees can share.

The company framed the agent around everyday business questions such as why sales slowed, where spending is rising, and which issues threaten renewals in the largest accounts — answers that, OpenAI noted, often require waiting for a report or asking someone else to run the analysis. Users direct and refine the analysis in a single conversation, without writing queries or learning a separate analytics tool, OpenAI said.

Data Connections and Business Context

The Data agent connects to approved data sources including Amazon Redshift, Datadog, Google BigQuery, ClickHouse, Databricks, MongoDB, and Snowflake, among others, and it can bring files and documents from Google Drive and SharePoint into an analysis.

To interpret that data, OpenAI said the agent applies an organization’s own business terms, metric definitions, custom calculations, and data relationships. That context comes from semantic layers and trusted sources such as Databricks Genie Ontology, dbt, GitHub, Snowflake Horizon, and existing business-intelligence dashboards.

Enterprise administrators choose which data connections are available and which roles can use them, and queries enforce the connected account’s existing permissions, including table-, row-, and column-level restrictions. OpenAI’s product page for data teams describes the same governance model, pairing role-based access control with citations and reviews for generated outputs.

The announcement includes statements from partners describing their own integrations, including AWS, ClickHouse, Databricks, Snowflake, MongoDB, and G2. MongoDB’s chief product officer for AI and emerging products, Pablo Stern, said the agent can reach data stored in Atlas as it updates in real time, while Snowflake’s head of developer experiences, Umesh Unnikrishnan, said employees can draw on Snowflake data under their existing access controls.

Dashboards, BI Tools, and Follow-Up Actions

An analysis can be turned into an interactive dashboard with built-in visualizations that teammates can edit, share, and refresh, and users can supply brand guidelines so outputs match the organization’s look and feel. The announcement shows an example adoption dashboard for a fictional “Acme Cloud” product covering June 2 to August 21, 2026, with sections for installs, activation, activity, and retention, generated from a prompt asking the agent to track adoption from installation and first use through ongoing activity and to identify where customers get stuck.

OpenAI said the agent can also build and interact with dashboards inside Omni, Oracle BI, Power BI, Sigma, Tableau, and ThoughtSpot, and the announcement carries statements from Tableau, Microsoft, Sigma, and ThoughtSpot describing those integrations. Bogdan Crivat, corporate vice president of Microsoft Fabric, said users can create Power BI dashboards by describing their business questions.

Beyond analysis and dashboards, OpenAI said ChatGPT Work can recommend next steps, identify who needs to be involved, share findings through Slack or email, and carry out user-approved actions through connected tools.

Adoption Inside OpenAI and Alpha Customers

OpenAI said nearly all of its product team and more than two-thirds of its go-to-market organization already use data agents in ChatGPT Work, adoption the company attributes to an internal data team that created shared business definitions, set access rules, and put safeguards in place for sensitive data.

Externally, OpenAI said NTT Data, Thermo Fisher, ServiceTitan, and other organizations in its Alpha program are using the agent to analyze sales and spending, catch reporting errors, and decide which opportunities to pursue and how to staff them. Yuji Shono, head of NTT DATA Group’s Global AI Office, said licensing costs, effort, and technical expertise had made it difficult to expand dashboards across his organization, adding that with the agent “many non-engineers, particularly in sales and corporate functions, have been able to build and update their own dashboards using plain language,” according to OpenAI’s announcement.

Ankur Bhatt, ServiceTitan’s vice president of engineering for Agent OS and applied AI, said his team used the agent to build a dashboard showing that users of Atlas, the company’s AI sidekick, launched campaigns at roughly three times the rate of nonusers, a finding he said is helping simplify onboarding. Other named alpha customers with statements in the announcement include Zipline, Empower, Piston, Doeren Mayhew, CookUnity, Turing, micro1, and Unit8. Zipline co-founder Ryan Oksenhorn said early testing surfaced findings that would otherwise have taken one of the company’s best people hours of digging to uncover, and micro1 strategic project lead Carlos Arroyo said his operations team rebuilt its performance-tracking dashboards in half an hour while catching errors in the originals.

Availability and Getting Started

The Data agent is listed as “Data” in the Plugins directory in ChatGPT Work. Administrators can install it or make it available to teams through Workspace settings, where they can also enable and configure data-source plugins such as Databricks and Snowflake and manage who can use them. Individual users install the plugin from the directory, complete the required account-connection steps, and start a conversation with @Data.

OpenAI published example prompts for the agent, including diagnosing why weekly active users changed, designing a KPI framework with primary metrics, drivers, guardrails, targets, and data-validation needs, and turning a month’s metrics into a leadership-ready update with actuals, comparisons, drivers, caveats, and recommended actions.

OpenAI also scheduled a webinar for September 10, 2026 at 9:30 AM PT on how its data team uses ChatGPT Work to turn data and business context into clear next steps.