Businesses generate enormous amounts of information every day, but finding useful answers inside that data can still take alot of hours. The new OpenAI Data Agent aims to change that by allowing employees to investigate company data, uncover trends, create interactive dashboards, and ask follow-up questions using natural language. The launch comes as OpenAI continues pushing AI beyond traditional chatbots and toward tools that can work directly with business systems, data, and workflows. For a broader look at this shift, see how AI is moving beyond the chatbot toward corporate ROI.
What Is the OpenAI Data Agent?
The OpenAI Data Agent is a data-analysis capability inside ChatGPT Work that helps people investigate business information without needing to write complex database queries or learn a specialized analytics platform.
Instead of submitting a request to a data team and waiting for a report, an employee can describe what they want to understand.
For example, a sales manager might ask why revenue declined during a particular period. A finance employee could investigate where spending is increasing. A customer-success team could look for changes in account retention.
The system can analyze connected data, explain what changed, provide evidence for its findings, and help transform the results into an interactive dashboard.
That makes data analysis more conversational. Employees can start with a business question rather than starting with a database.
How the ChatGPT Data Agent Works
The ChatGPT Data Agent is designed to work with the information and business context an organization already trusts.
OpenAI said Data can connect with approved enterprise sources such as Amazon Redshift, ClickHouse, Databricks, Google BigQuery, MongoDB, and Snowflake. It can also work with files and documents from services, including Google Drive and SharePoint when those connections are available.
This means companies do not necessarily have to abandon their existing data infrastructure to use the technology.
The system can investigate information from connected sources and use the organization’s existing definitions to understand what different metrics mean.
That is particularly important for businesses where terms such as revenue, active customers, retention, conversion, or qualified leads have specific internal definitions.
Why Business Context Matters
An AI system can produce an impressive answer and still provide the wrong conclusion if it misunderstands the data behind a business metric.
OpenAI says Data can use semantic layers and trusted business sources to understand metric definitions, custom calculations, and relationships between datasets.
This allows the OpenAI Data Agent to work with the terminology already used by an organization rather than treating every dataset as if it were isolated.
For example, if a company has a specific definition for customer retention, that definition can provide important context when the system investigates why retention changed.
The same principle applies to dashboards and existing business intelligence systems.
The objective is not simply to make AI capable of reading data. It is to make the analysis more consistent with the way a company already understands its business.
Turn Questions Into Interactive Dashboards
One of the most useful features is the ability to turn analysis into an interactive dashboard.
After asking a question, users can continue the conversation to investigate specific findings, compare different groups, examine changes over time, or review the evidence supporting an answer.
They can then ask the system to turn the analysis into a dashboard containing visualizations.
The dashboard can be edited, shared, and refreshed as necessary.
Businesses can also provide brand guidelines to customize the appearance of their outputs.
This could make sophisticated analytics more accessible to employees who understand their business but do not have advanced data-engineering or business-intelligence skills.
ChatGPT Work Connects Data With Existing BI Tools
The new capability is not positioned as a replacement for every analytics platform.
OpenAI says ChatGPT Work can also build and interact with dashboards in tools including Omni, Oracle BI, Microsoft Power BI, Sigma, Tableau, and ThoughtSpot.
That could be useful for organizations that have already invested in business intelligence.
Rather than forcing employees to learn an entirely new analytics workflow, the ChatGPT Data Agent can provide a conversational way to investigate questions and work with information that eventually feeds into existing business intelligence environments.
For a marketing manager, this could mean asking which campaigns are producing the strongest results.
For a finance team, it could mean investigating unusual spending.
For an operations manager, it could mean identifying the biggest changes in performance.
The starting point is simply the question.
From Data Analysis to Action
The bigger idea behind the product is that analysis does not have to end with a report.
The OpenAI Data Agent can help users determine potential next steps and identify people who may need to be involved.
OpenAI also says findings can be shared through connected workplace tools such as Slack or email, while approved actions can be carried out through connected systems.
That creates a workflow that moves from:
Question → Analysis → Evidence → Dashboard → Recommendation → Action
This is different from the traditional process where an employee asks another team for a report, waits for the report, reviews it, and then submits another request when a new question appears.
The goal is to shorten that cycle.
OpenAI Is Using Data Agents Internally
OpenAI says it already uses the capabilities behind the Data agent across its own organization.
The company says nearly all of its product team and more than two-thirds of its go-to-market organization use data agents in ChatGPT Work to analyze company information.
According to OpenAI, its internal implementation required shared business definitions, access controls, and safeguards for sensitive information.
This is significant because it shows the technology is being positioned as an everyday business tool rather than simply an experimental AI feature.
The idea is to allow more employees to perform useful analysis themselves while specialized data teams can concentrate on more complex problems.
Companies Are Testing the OpenAI Data Agent
Organizations participating in OpenAI’s Alpha program are already testing the technology for different business tasks. NTT DATA says non-engineering employees have used the system to create and update dashboards through natural-language requests. Thermo Fisher Scientific has been testing it with existing company data to help teams prepare more effectively and identify opportunities involving its supply base. ServiceTitan used the technology to build a dashboard examining campaign activity among users of its Atlas AI product. Other organizations have reportedly used it to examine sales performance, spending, customer satisfaction, operational metrics, adoption, forecasting, and business development. These examples show how broad the potential use cases can be. The technology does not have to belong exclusively to a company’s analytics department.
Security and Data Access
Giving an AI system access to company information naturally raises security and privacy questions.
OpenAI says workspace administrators control which connections are available and which roles can use them.
Queries also respect permissions from connected accounts, including applicable table, row, and column restrictions.
That means companies can determine which employees are allowed to access particular data sources.
This type of governance is becoming increasingly important as AI systems gain access to more enterprise information.
The more useful AI becomes inside a company, the more important it becomes to control what the system can access and what employees can do with the resulting information.
That concern is also relevant to OpenAI’s broader development of increasingly capable AI systems. Our coverage of OpenAI’s cybersecurity capabilities and the future of AI safety looks at that wider issue.
Why the OpenAI Data Agent Could Be Important for Businesses
Large companies often have dedicated data analysts, engineers, financial teams, and business-intelligence specialists.
Smaller businesses may have far fewer people available to perform that work.
A conversational data tool could therefore have a particularly significant effect on organizations with limited analytical resources.
A business owner could investigate sales performance.
A marketing manager could examine campaign results.
A finance employee could look for unusual spending patterns.
An operations manager could investigate why productivity has changed.
The value is not necessarily that AI replaces data professionals. Instead, it can help more employees answer straightforward questions independently and reserve specialist expertise for more complicated analytical work.
AI Is Moving Into Business Operations
The arrival of the OpenAI Data Agent is part of a larger movement in enterprise AI. Early workplace AI tools were primarily used for writing, summarizing, brainstorming, coding, and answering questions. Increasingly, businesses want AI systems that can work with the applications and information they already use. That includes customer databases, financial systems, analytics platforms, internal documents, communications tools, and software-development environments.
OpenAI’s expansion into different markets reflects the same broader trend. Its growing presence in Brazil comes as organizations increasingly explore practical applications for AI. Read more about OpenAI’s expansion in Brazil and the rise of AI adoption. The Data agent fits into this transition by putting business data closer to the point where decisions are made.
How to Use the ChatGPT Data Agent
Organizations can make data available through the Plugins directory in ChatGPT Work.
Workspace administrators can control whether the capability is available and configure the data-source connections required by their teams.
Once the necessary connections have been configured, users can begin asking business questions through the Data experience.
Examples include
- Why did revenue fall this quarter?
- Which regions are growing fastest?
- What caused customer retention to decline?
- Build a dashboard showing sales performance by region.
- Compare this month’s spending with the previous month.
- Which accounts require attention?
- What changed in our operational performance?
OpenAI’s current documentation similarly recommends specifying the source, metric, time period, and comparison when possible so the analysis has enough context.
What the OpenAI Data Agent Means for the Future
The OpenAI Data Agent could make business analytics more accessible by reducing the technical barrier between employees and company information.
It does not remove the need for human judgment. Important findings still need to be checked against their sources, definitions, time periods, and filters before they are used to make significant decisions.
But the ability to move from a business question to analysis and then to an interactive dashboard could significantly shorten the path from information to action.
The bigger change is cultural as much as technical. Data may no longer have to be something employees request from a specialized department and receive later as a static report. Instead, more people can potentially ask questions directly, investigate the evidence, explore the results, and decide what should happen next. As businesses connect more of their systems to AI, the advantage may increasingly belong to organizations that can turn trusted data into useful decisions quickly. And that is where the OpenAI Data Agent could become one of the more important developments in the next generation of workplace AI.
