OpenAI launched ChatGPT for Financial Services on September 10, 2026, a version of ChatGPT Work tailored for investment banks and based on GPT-6 Astra. Morgan Stanley and Evercore assisted as design partners to focus the product on research, financial models and pitchbooks. Junior bankers, who often work hundred-hour weeks, are meant to benefit most.
Data from multiple providers flows directly into the chat
The new offering builds on ChatGPT Work and connects the chat with financial data from multiple providers. According to OpenAI, built-in sources include Daloopa, PitchBook and LSEG News, so analysts do not need separate licensing agreements for them. Anyone who already has access to S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva or Moody’s can connect these through their own accounts; optimized connectors to S&P Global and FactSet round out more than fifty data connectors in total.
The system links figures directly back to the underlying filing or report, and charts can be checked against the original data. Permissions are meant to stop confidential deal documents from becoming visible to the wrong teams by accident — a point that matters to investment banks handling insider information. The tool also produces ready-made Excel, Word and PowerPoint templates, for pitchbooks or research notes. OpenAI discloses no pricing and no launch date for Germany or the EU. For now, only selected institutions get access through OpenAI’s own account team; there is no public self-service tier.
Nick Turley points to bankers’ hundred-hour weeks
OpenAI product chief Nick Turley reportedly wants the model to research like an analyst and back up its conclusions the same way. According to CNBC, Turley cited junior bankers’ frequent hundred-hour weeks as the starting point for the development. Asked whether the tool would mean fewer new hires at investment banks, he reportedly sidestepped the question and spoke instead of more output per employee.
Morgan Stanley and Evercore, as design partners, shaped the focus on investment banking and equity research, where reliable data access and high-quality documents have been the biggest time sinks, according to OpenAI. Both firms voiced cautious optimism about the project without giving concrete figures on time savings or cost. Beyond the two pilot partners, OpenAI stayed quiet on which other banks are testing the product or when broader availability might follow.
OpenAI catches up to Anthropic’s head start in finance
With the product, OpenAI is challenging an already established rival: Anthropic launched Claude for Financial Services back in July 2025, integrating Snowflake, S&P Global and Morningstar among others. Clients such as the hedge fund Bridgewater and Norway’s sovereign wealth fund reportedly already use the Anthropic version in production. OpenAI enters the same market just over a year later, relying on its own set of partnerships rather than identical data providers.
The move fits OpenAI’s broader push into consulting and enterprise accounts: just in August, IBM equipped its consulting arm with OpenAI models and explicitly named financial services as a target industry. For investment banks, the announcement mainly signals one thing: competition for AI tools in research and modeling is intensifying, while individual vendors disclose few comparable figures on impact or cost.
What will matter is whether ChatGPT for Financial Services wins major banks beyond Morgan Stanley and Evercore, while Anthropic can already point to reference clients there. It also remains open whether the promised efficiency actually cuts junior bankers’ working hours or merely shifts their tasks — a question that will only be answered once the first institutions report publicly on their experience.
