OpenAI is making a strategic move into institutional finance with a specialized AI assistant. On September 10, 2026, the company launched ChatGPT for Financial Services, an optimized version of its conversational AI targeting investment banking and equity research professionals. Built on the GPT-6 Astra model in collaboration with Morgan Stanley and advisory firm Evercore, this product aims to transform workflows at major financial institutions.
🔑 Key Takeaways
- GPT-6 Astra model developed with Morgan Stanley and Evercore
- Primary targets: investment banking and equity research teams
- Approximately 50 MCP connectors to Bloomberg, FactSet, Crunchbase and PitchBook
- Production of pitch-books and financial models in approximately 10 minutes
An AI Assistant Tailored for Institutional Finance
ChatGPT for Financial Services positions itself as a tool capable of researching companies, analyzing financial data, and generating PowerPoint presentations, Excel spreadsheets, and web dashboards. It can also produce adaptive financial models and pitch-books using banks’ internal templates. During a demonstration, the system evaluated a potential acquisition target, extracted financial figures from standard industry sources, and produced a PowerPoint deck formatted according to a bank’s style guide, all in approximately ten minutes.
The interface resembles standard ChatGPT but includes toggles for integrating financial data from providers such as Bloomberg and FactSet. OpenAI also offers pre-loaded datasets from Crunchbase, PitchBook, Daloopa, and LSEG News. In total, approximately 50 MCP (Model Context Protocol) connectors enable integration with other software and data sources.
« Each time the model uses this data, it provides detailed citations allowing users to trace the origin of information. Graph audit tools also allow users to verify the consistency of visualizations with underlying data. »
OpenAI, product documentation
Features and Performance Tiers
Users can select the effort level deployed by the model: high, medium, or low. A high effort level consumes more tokens and responds more slowly but generally produces higher quality output. This mechanism aims to match performance to the specific requirements of each task.
| Level | Quality | Speed | Use Cases |
|---|---|---|---|
| Low | Standard | Fast | Simple queries, verification |
| Medium | Balanced | Moderate | Standard analysis, reports |
| High | Optimal | Slower | Complex models, pitch-books |
Enterprise Security and Control
Regarding security and confidentiality, OpenAI states that data is encrypted and protected by SAML SSO protocols, SCIM provisioning, and role-based access controls specific to ChatGPT Enterprise. Administrators can configure data retention policies and define which capabilities are accessible to each role within the organization.
« We’re trying to think of new ways of doing the work, rather than just making the existing ways faster. »
Joseph Kim, OpenAI Product Lead
Market Positioning and Industry Reactions
The product primarily targets investment banking and equity research teams. It enables preparation of market updates, potential buyer lists, acquisition/dilution models, and other client documents. During the briefing, Nick Turley stated: « This is the canonical product we are hoping the industry adopts. » He also emphasized that the goal is to make ChatGPT for Financial Services the only tool needed by large banks or financial companies employing tens of thousands of employees.
« There’s a difference between what looks good in a demo and what is actually a usable output, [and] you kind of rely on the experts to achieve that. »
Nick Turley, VP ChatGPT, OpenAI
On the competitive landscape, OpenAI is not the first in this segment. Anthropic launched Claude for Financial Analysis in May or July 2025 depending on sources. OpenAI believes its offering will stand out through data integration depth and citation quality. Turley acknowledged strong demand (« there’s a ton of demand ») but declined to name institutions that have already subscribed.
Implications and Outlook for the Profession
The arrival of this tool has raised questions about the future of junior banker roles. Chris Churchman, partner at Goldman Sachs and leader of one of the bank’s major AI initiatives, warned about the risk of « cognitive atrophy » if tasks traditionally assigned to analysts for training purposes are automated. « Reasoning is still important, » he stated. « You still need to reason about problems and structure it into an argument, and now we’re delegating reasoning. »
« Reasoning is still important. You still need to reason about problems and structure it into an argument, and now we’re delegating reasoning. »
Chris Churchman, Partner, Goldman Sachs
Turley responded that the product is primarily a productivity lever, comparing its impact to Microsoft Excel, which allowed the industry to produce analyses more quickly. From a financial perspective, Sarah Friar, CFO of OpenAI, stated in August 2026 that enterprise revenue now exceeded consumer business revenue. OpenAI is also preparing for a highly anticipated IPO as a major milestone for the company.
Sources
This article is published for informational and educational purposes. It does not constitute investment advice. Conduct your own research (DYOR) before making any decisions.
