Financial Data Moves Inside ChatGPT
ChatGPT for Financial Services includes datasets from Daloopa, PitchBook, LSEG News and Crunchbase. The available information covers areas including earnings transcripts, financial statements, company fundamentals, private businesses, funding activity and acquisitions.
OpenAI said these built-in datasets are indexed and hosted on its infrastructure, allowing the platform to retrieve information more quickly and provide granular citations. An analyst examining a company’s adjusted earnings, for example, would be able to trace a figure to the supporting table or passage instead of receiving an answer without a visible source trail.
Financial institutions can also connect data covered by their existing subscriptions. OpenAI is working with providers including S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva and Moody’s on shared sign-in and entitlement integrations. Its broader connector network includes services such as FactSet, Preqin, Datasite, Box and Intapp.
Sally Moore, chief client officer and co-head of Market Intelligence at Kensho Data & Platforms, S&P Global, said: “Trusted data is the foundation every AI workflow is built on. Bringing the S&P Global AI Data Portal to ChatGPT gives financial professionals the full breadth of our verified intelligence, from financials and transcripts to market and energy data, right where they’re already working. We’re proud to collaborate with OpenAI to deliver Adaptive and Deterministic Retrieval through ChatGPT for Financial Services.”
The distinction between included data and subscription-based access is important to understanding the product. Daloopa, PitchBook, LSEG News and Crunchbase are presented as built-in sources, while some other services require the financial institution to maintain the relevant commercial subscription. OpenAI has not suggested that every external dataset will be available to every user automatically.
Emily Prince, LSEG’s group head of enterprise AI, said: “Our collaboration with OpenAI advances our LSEG Everywhere AI strategy. With MCP connectivity and a curated Reuters news selection, we’re focused on making LSEG’s trusted, licensed content and analytics available to customers wherever they want to work.”
From Company Research to Client Materials
Research is only one part of the offering. OpenAI says the system can compare companies, analyze earnings, test valuation assumptions and produce editable financial models. It can also synthesize the results into spreadsheets, written research and presentation materials.
Administrators can publish approved Excel, Word and PowerPoint templates through a dedicated management page. Once a firm’s templates and style guidance are configured, employees can use them to prepare valuation models, research notes and pitchbooks in the organization’s customary format.
This approach addresses a routine but time-consuming part of financial work. Analysts often move repeatedly between data terminals, spreadsheets, internal documents and presentation software. Keeping those materials connected could reduce manual transfers, although firms will still need to review the underlying data, formulas and conclusions before using the output in client work.
The product runs on GPT-6 Astra, which OpenAI describes as being designed for information retrieval, financial reasoning and the generation of documents and other work products. The model can interpret figures, tables and accompanying notes, conduct an analysis and transfer the results into editable documents, spreadsheets or slides.
OpenAI has published performance results for the model, including a score of 69.9% on OfficeQA Pro, an evaluation involving complex US Treasury documents, tables and footnotes. GPT-5.6 Sol scored 60.2% on the same benchmark.
Controls for Regulated Institutions
Handling confidential client information and material nonpublic information presents a higher threshold than using a public AI chatbot for general research. OpenAI has therefore based the new service on the security and governance controls available through ChatGPT Enterprise.
Those controls include SAML single sign-on, automated user provisioning and role-based access. OpenAI says customers’ business data is not used to train its models by default and is encrypted both while stored and while being transmitted.
Administrators can set data-retention rules, manage access to connected applications and restrict supported read or write actions according to an employee’s role. Institutions may also establish separate workspaces to support information barriers between teams.
Compliance departments can export supported workspace logs through the OpenAI Compliance Platform for use in existing audit and investigation processes. These features are intended to give institutions more visibility over how employees access data and use AI-generated material, though each firm remains responsible for determining whether its deployment satisfies internal policies and applicable regulation.
OpenAI has not publicly disclosed pricing or provided a timetable for broad availability. The company says ChatGPT for Financial Services is available to “eligible financial institutions,” which must contact OpenAI or their account representative to seek access.
