On Thursday, OpenAI launched ChatGPT for Financial Services, a specialized ChatGPT experience designed to help financial institutions conduct research, develop financial models and create custom content for clients.
The product combines embedded financial data with Astra’s GPT-6 reasoning capabilities and was developed in collaboration with Morgan Stanley and Evercore. OpenAI said the collaboration helped identify the biggest challenges facing financial institutions and shape solutions aimed at supporting bankers in their daily work.
What financial data does ChatGPT offer for financial services?
ChatGPT for Financial Services includes datasets from providers such as Daloopa, PitchBook, LSEG News and Crunchbase. They cover areas such as earnings breakdowns, financial statements, company fundamentals and private companies.
OpenAI said the citations accompanying the data allow bankers to trace figures and claims back to their original sources and check the evidence to support their analysis. According to OpenAI, teams can start working with data sets without having to negotiate separate contracts or install separate connectors. The company said it indexes and hosts data on its own infrastructure, which improves retrieval, latency and the way information is displayed in the product.
For example, a banker conducting an earnings normalization analysis may review the reconciliation and notes to the adjusted Ebitda figure, understand what costs have been eliminated, and then decide how the adjusted figure should be used in the valuation.
How does GPT-6 Astra help in financial analysis?
The product also offers advanced OpenAI intelligence, including GPT-6 Astra. OpenAI said new models will continue to be available through the product as they are released.
GPT-6 Astra is designed to support three core functions required to operate financial services: information retrieval, financial justification and artifact generation. The model can navigate and understand numbers, tables, and accompanying notes in financial documents. It can take inputs, perform financial analysis and draw conclusions, and then synthesize the results into documents, tables, and slides.
OpenAI said it has built capabilities into its product that allow teams to conduct research across multiple sources, track numbers over different periods, and interpret annotations in publicly available data. Once the analysis is complete, teams can also create interactive charts and visualizations.
How can fintechs adapt issuance?
ChatGPT for Financial Services allows admins to publish Excel, Word and PowerPoint templates through a dedicated admin page.
By customizing firm templates and style guides, teams can turn their analysis into assessment models, research notes, and books using their firm’s preferred format and style. The goal is to enable financial institutions to move from research and analysis to ready-made client or internal content without having to recreate the experience in separate applications.
How can companies access financial data subscriptions?
OpenAI said it recognizes that many customers already have subscriptions to financial data providers. Therefore, the company is working with S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva and Moody’s to integrate common entry and rights.
Under these agreements, providers will be able to recognize users by logging into ChatGPT and automatically grant access to data to which those users already have rights.
How will companies’ financial data be protected?
OpenAI said business data from firms is not used to train its models by default. Data is encrypted both at rest and in transit, and administrators can configure workspace storage options.
Compliance teams can export supported workspace logs through the OpenAI Compliance Platform into existing audit and investigation workflows. The product is also based on ChatGPT Enterprise’s security and governance controls, including SAML single sign-on (SSO), SCIM provisioning, and role-based access control.
Companies can control access to skills and applications based on employee roles, and enable or disable supported application read and write actions. OpenAI said organizations can also create multiple workspaces to strengthen information barriers between teams or groups within a firm.