OpenAI Launches Data Agent in ChatGPT Work for Business Analytics
OpenAI's new Data agent in ChatGPT Work connects to company data, enabling real-time insights and interactive dashboards via natural language.
OpenAI has unveiled the Data agent within ChatGPT Work, an advanced tool designed to simplify data analysis and visualization for businesses. Released on September 9, 2026, the Data agent enables users to connect directly to their company’s data sources, ask questions in natural language, and generate interactive dashboards without requiring technical expertise.
This launch builds on OpenAI’s broader push into enterprise tools through ChatGPT Work, a business-oriented mode of ChatGPT focused on complex, multi-step tasks. The Data agent represents a significant step forward, targeting a common bottleneck in organizations: the reliance on technical teams for data insights and reporting.
The Data agent integrates with major enterprise data platforms, including Amazon Redshift, Snowflake, Google BigQuery, Databricks, and MongoDB. It also pulls in files from Google Drive and SharePoint for analysis. By utilizing semantic layers and business-specific contexts—like metric definitions and data relationships from tools like dbt and Databricks Genie Ontology—the agent ensures results are aligned with the organization’s established data standards.
“With the Data agent in ChatGPT Work, users can connect directly to Amazon Redshift, making it as easy as asking a question to explore data, surface insights, and make faster decisions across their organization,” said Naresh Chainani, Director of Engineering at AWS.
Users can interact with the Data agent in plain language to investigate trends, diagnose anomalies, and refine analysis in real time. The findings can then be converted into shareable, interactive dashboards with built-in visualizations. Compatible tools include Tableau, Power BI, Oracle BI, and ThoughtSpot, allowing seamless integration with existing workflows. Outputs can also be customized to match brand guidelines.
Southard Jones, EVP & Chief Product Officer at Tableau, highlighted the value of this integration: “Connecting Tableau with ChatGPT Work brings trusted business semantics into a place employees already work, so the answers they get are grounded in the same data model their teams rely on.”
Administrators retain control over which data sources are accessible and by whom. The tool enforces existing permissions, ensuring compliance with table, row, and column-level restrictions. This focus on security and governance aligns with OpenAI’s ongoing efforts to position ChatGPT Work as a reliable enterprise solution.
The Data agent is already in use by companies like NTT Data, Thermo Fisher, and ServiceTitan. Yuji Shono, Head of Global AI Office at NTT Data Group, noted, “With the Data agent in ChatGPT Work, many non-engineers, particularly in sales and corporate functions, have been able to build and update their own dashboards using plain language.”
This launch comes as OpenAI accelerates its shift toward enterprise tools to compete with rivals like Anthropic. Recent releases, such as GPT-5.6 and other ChatGPT Work capabilities, underline its strategy to dominate the business-focused AI market.
The Data agent is available in the Plugins directory of ChatGPT Work. Administrators can enable it via Workspace settings, configure data-source connections like Databricks or Snowflake, and manage user permissions. For a hands-on experience, users can directly converse with the agent to ask business questions and explore data.
By democratizing access to data analytics, OpenAI is positioning the Data agent as a critical tool for businesses aiming to make faster, data-driven decisions without the traditional technical barriers.
Asanat Analysis — Why it matters
OpenAI's Data agent targets a structural inefficiency in enterprise operations: the gatekeeping of data insights by technical teams. By enabling natural-language querying across Snowflake, Redshift, BigQuery, and similar platforms, the tool lowers friction for non-technical users to access analytics—a shift that historically required either upskilling or hiring. The integration with semantic layers (dbt, Databricks Genie) signals OpenAI's awareness that LLM outputs require grounding in organizational data models to avoid hallucination or misinterpretation of metrics, a critical requirement for business adoption.
The vendor ecosystem response—from Tableau, Salesforce-adjacent BI players, and cloud infrastructure providers—indicates this is perceived as a genuine competitive threat to existing analytics workflows, not an experimental feature. AWS and Databricks co-promoting the integration suggests these platforms see LLM-mediated analytics as table-stakes for enterprise stickiness. The permission enforcement mechanics (table/row/column-level access control) reveal OpenAI's focus on compliance requirements, a prerequisite for serious enterprise deployment that contrasts with earlier consumer-oriented releases.
For crypto/DeFi, this precedent matters indirectly: if LLM-based analytics become the standard for institutional data consumption, decentralized analytics platforms and on-chain analytics tools (Dune, Flipside) face pressure to integrate similar UX patterns or risk obsolescence to centralized incumbents. Early-stage on-chain analytics startups should monitor whether large language models can reliably operate on blockchain data semantics without losing the deterministic guarantees that currently differentiate crypto analytics from traditional BI.