OpenAI Launches ChatGPT for Finance with Integrated Data

OpenAI has introduced a version of ChatGPT tailored for financial services, built on GPT-6 Astra and premium datasets. Designed in collaboration with Morgan Stanley and Evercore, the tool combines traceability of figures, multiple connectors, and centralized administration. Integrations with major financial information providers are in the works.

Ready-to-Use Deliverables and Traceable Analyses

The product allows users to query multiple sources, reconcile figures over different periods, and interpret annotations derived from public data. After analysis, teams can generate interactive charts and visualizations. Administrators have a dedicated page to publish Excel, Word, and PowerPoint templates for users. Figures and claims can be linked to specific tables and passages, with supporting elements highlighted. During a P&L normalization analysis, a banker can examine the reconciliation and notes associated with adjusted EBITDA, understand excluded costs, and decide whether to include them in an evaluation. GPT-6 Astra is utilized to comprehend tables and notes, conduct financial analyses from structured inputs, and produce documents, spreadsheets, and summary presentations.

Data Access, Simplified Contracts, and Connectors

Datasets from Daloopa, PitchBook, LSEG News, and Crunchbase cover earnings call transcripts, financial statements, company fundamentals, and private companies. Teams can access these immediately, without additional contracts or connector setup, as the data is indexed and hosted by OpenAI to enhance retrieval, latency, and presentation, with granular citations and tracking back to sources. OpenAI is collaborating with S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva, and Moody’s to enable connection integrations and shared rights, allowing providers to recognize users via ChatGPT login and activate automatic access to already subscribed data. In response to the challenges of using certain MCP connectors, OpenAI has optimized widely used connectors for immediate use and lists over 50 integrations, including Datasite, Box, Preqin, and Intapp.

Enterprise Controls and Product Roadmap

Organizations can centrally manage access and data links, relying on ChatGPT's security and governance mechanisms. OpenAI plans to add new models as they become available and guide post-training and functional developments with its partners. The company aims to expand data coverage and train its models to find, interpret, and leverage this content with a level of expertise comparable to that of top analysts.

Initial Framing by Investment Banking and Equity Research

The initial use cases prioritize investment banking and equity research, following work conducted with Morgan Stanley and Evercore. According to OpenAI, reliable access to data and the production of quality artifacts were among the main challenges for these teams. Morgan Stanley explains that it is working with OpenAI to integrate these capabilities into corporate research, analysis development, and advisory preparation, testing ideas and contributing to design. The bank describes a product gradually informed by its methods and teams. Evercore indicates that it is closely collaborating to deepen the insights underlying its recommendations, relying on its judgment and the standards expected by its clients.

Technical Foundation: GPT-6 Astra and Hosted Data

ChatGPT for finance combines integrated data and GPT-6 Astra to assist in research, modeling, and support production. The model is presented as cutting-edge in information retrieval, financial reasoning, and artifact generation. The service aggregates financial data with the expected level of detail, including premium datasets and streamlined connections, and announces improvements in the performance of MCP connectors.