SAN FRANCISCO — OpenAI is launching a version of ChatGPT tailored specifically for financial institutions, combining its latest artificial intelligence technology with built-in premium financial data and tools for producing research, financial models, spreadsheets and client materials.
The company said ChatGPT for Financial Services is a specialized ChatGPT Work experience powered by GPT-6 Astra and initially focused on investment banking and equity research.
OpenAI developed the product in partnership with Morgan Stanley and Evercore, which helped identify some of the most significant challenges financial institutions face when incorporating AI into bankers’ daily workflows.
What’s Included
The new service includes financial data from providers including Daloopa, PitchBook and LSEG News, covering areas such as earnings transcripts, financial statements, company fundamentals and private companies. OpenAI said the data is indexed and hosted on its infrastructure, allowing the company to improve retrieval speeds and provide detailed citations that enable users to trace figures and claims to their original sources.
OpenAI said the approach also eliminates the need for institutions to separately negotiate contracts or configure connectors for data included with the service.
Existing Data Subscriptions Can Also Be Connected
OpenAI said it is also working with S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva and Moody’s on integrations that will recognize users through their ChatGPT sign-in and provide access to data for which their institutions already have subscriptions.
The company’s broader ecosystem includes more than 50 connectors, including services from Datasite, Box, Preqin, FactSet and Intapp.
OpenAI said it has also focused on improving the performance and reliability of Model Context Protocol, or MCP, connections commonly used by financial institutions to access external data.
The company said the combination is intended to reduce the time financial professionals spend locating and connecting data and allow them to focus more heavily on analysis.
For example, a banker conducting a profit-and-loss normalization analysis could examine the reconciliation and supporting notes behind adjusted EBITDA, identify costs excluded from the calculation and determine how those adjustments should be treated in a valuation.
GPT-6 Astra at Center of Offering
The product incorporates GPT-6 Astra, OpenAI’s newest model for professional work.
OpenAI said the model is designed to perform three capabilities particularly important in financial services: information retrieval, financial reasoning and creation of finished work products.
The model can analyze figures, tables and supporting notes in financial documents, perform financial analysis and synthesize the results into documents, spreadsheets and presentations, according to the company.
ChatGPT for Financial Services can also research across multiple sources, trace financial figures across different reporting periods and interpret annotations accompanying public financial data. Users can then turn the analysis into interactive charts and other visualizations.
OpenAI said the product is designed to address two issues identified through its work with Morgan Stanley and Evercore: reliable access to financial data and the ability to turn analysis into high-quality finished materials.
Firm Templates Can Be Built In
Financial institutions can also configure ChatGPT to produce work using their own templates and style standards.
Administrators can publish Excel, Word and PowerPoint templates through a centralized administration page. Employees can then use those templates to create materials such as valuation models, equity research reports and investment-banking pitchbooks in their firm’s established formats.
OpenAI said it plans to use feedback from financial-services partners to improve the product, further train its models to work with specialized financial datasets and eventually expand into additional areas of financial services.
Security, Governance Emphasized
Because financial institutions routinely handle material nonpublic information and confidential client data, OpenAI said ChatGPT for Financial Services includes enterprise security and governance controls.
The service builds on ChatGPT Enterprise capabilities including SAML single sign-on, SCIM provisioning and role-based access controls.
OpenAI said customers’ business data is not used to train its models by default and is encrypted both at rest and in transit. Administrators can also establish workspace retention policies.
Compliance teams can export supported workspace logs through the OpenAI Compliance Platform for use in existing audit and investigation processes.
Institutions Can Control Access
Institutions can also control access to skills and apps by employee role and determine whether supported apps are permitted to perform read or write actions. OpenAI said institutions can create multiple workspaces to establish information barriers within an organization.
The company said financial institutions can also use OpenAI’s application programming interface to build specialized applications for areas such as research, financial analysis, operations and other industry-specific needs.
ChatGPT for Financial Services is available to eligible financial institutions, according to OpenAI.
