Inside ChatGPT for Financial Services: What early access users can actually do

OpenAI has introduced a tailored version of ChatGPT intended for professionals in the financial sector. This specialized iteration, known as ChatGPT for Financial Services, integrates GPT-6 Astra technology with financial data, research tools, and the capability to construct valuation models and presentations. Unlike regular ChatGPT, which requires manual querying for financial information, this version houses relevant data within itself. Here are the features, eligible users, and how to initiate this product.

ChatGPT for Financial Services is essentially ChatGPT customized to mirror the processes of an investment banker or financial analyst. Initially developed in collaboration with Morgan Stanley and Evercore, the primary focus was on investment banking and equity research. The tool's functionality includes researching companies, analyzing earnings, comparing financial performance, constructing models, and converting work into various documents, spreadsheets, and presentations.

The distinguishing factor lies in the data it uses. The feature incorporates high-quality datasets from Daloopa, PitchBook, LSEG News, and Crunchbase, covering financial statements, earnings transcripts, company fundamentals, and private companies. However, to gain access to ChatGPT for Financial Services, one cannot simply enable it within their ChatGPT account.

This specialized version is presently accessible to eligible financial institutions. Interested firms must reach out to OpenAI or their existing OpenAI account team for initiation.

While individual ChatGPT subscribers have access to a personal finance feature, the financial version is not open for independent use. OpenAI is also collaborating with providers such as S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva, and Moody’s, allowing companies to utilize existing subscriptions within ChatGPT. Furthermore, over 50 connectors are available within OpenAI's broader ecosystem, including services like Datasite, Box, Preqin, FactSet, and Intapp.

This integration allows analysts to research across these sources from a single workspace, streamlining the process.

Upon obtaining the necessary data, users can assign tasks to GPT-6 Astra instead of merely posing questions. For instance, an analyst could instruct ChatGPT to compare a company's performance across multiple quarters, examine earnings data, or investigate the calculation of adjusted EBITDA. The model can trace figures through tables, documents, and supporting notes, subsequently utilizing this information in its analysis.

Importantly, it provides granular citations, enabling users to trace figures and claims back to specific tables and passages. However, it is essential to remember that this does not replace the need for human verification of AI-generated financial analysis.

Beyond research, ChatGPT for Financial Services can transform its findings into Excel spreadsheets, Word documents, and PowerPoint presentations. Administrators can also upload approved templates and style guides. Consequently, an analyst might potentially progress from researching a company to producing a valuation model, research note, or pitchbook without manually recreating the materials in the firm's standard format.

This feature could significantly save time for those who have previously spent hours transferring research into slides or spreadsheets.

Regarding security, ChatGPT for Financial Services builds upon ChatGPT Enterprise's existing safeguards, including SAML SSO, SCIM provisioning, and role-based access. OpenAI assures that business data is not used for model training by default and is encrypted both at rest and in transit. Administrators also have control over data retention, export supported workspace logs for compliance purposes, and the creation of separate workspaces to maintain information barriers.

Although this product is currently exclusively available to eligible financial institutions, its implications extend beyond the immediate financial sector. This development indicates that OpenAI is working towards integrating its models with specialized data, permissions, templates, and workflows designed for specific job requirements.

For the finance industry, this means GPT-6 Astra is positioned directly alongside financial databases, company information, and document formats, illustrating a broader trend in AI applications.