There is a particular kind of strategic audacity in helping a technology company build the product that automates your own workforce. That is precisely what Morgan Stanley and Evercore chose to do, and whether history judges it as brilliant or reckless depends almost entirely on what happens in the next five years at every major bank that did not get a seat at the table.
On September 10, OpenAI launched ChatGPT for Financial Services, a tailored ChatGPT Work experience that pairs built-in financial data with the reasoning of its GPT-6 Astra model to help teams develop research, financial models, and customized client materials. The product was shaped by a design partnership with Morgan Stanley and Evercore, whose early work steered the product toward investment banking and equity research, where reliable data access and high-quality artifact creation proved to be the sharpest pain points.
What the Product Actually Does
The tasks it performs out of the box: comparable-company analysis, LBO modeling, buyer screening, earnings analysis, pitchbook preparation. These are the tasks Wall Street has assigned to its youngest employees for decades. The product includes premium datasets from providers including Daloopa, PitchBook, LSEG News, and Crunchbase. For firms that already hold data subscriptions, OpenAI says it is developing shared sign-in and entitlement integrations with S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva, and Moody’s.
Other features tailored for finance include citations that allow users to trace data back to source filings and audit charts, as well as administrative controls for sensitive deal materials. The data friction that consumed analyst hours, chasing access, stitching together sources, is largely dissolved.
The Question No One at Morgan Stanley Is Answering Publicly
Why build it rather than buy it? The obvious read is defensive: if a frontier lab is going to automate your junior talent pool regardless, you might as well be the firm whose workflows, pain points, and institutional templates shape the product. OpenAI has said the partner work will inform post-training, product improvements, and its expansion into other financial services categories. That is not a vendor relationship. It is influence over the roadmap of the tool that will compete with your own headcount.
Morgan Stanley secured a significant first-mover advantage when it announced it was working with OpenAI around the period GPT-4 was released, and the collaboration has gone well beyond simple licensing. The design partnership on ChatGPT for Financial Services is an extension of that logic, carried further. Morgan Stanley is not just licensing a product. It is co-authoring one, embedding its own institutional knowledge into a model that every competitor will eventually use.
The risk is the mirror image. If the product ships to every bank at similar capability, the design advantage evaporates and what remains is a more efficient industry with fewer junior hires across the board. Some executives and practitioners have raised a related worry: automate too much of the apprenticeship work, and you risk dulling the on-the-job learning that turns analysts into sound, independent operators. That concern cuts across the whole industry. A bank that helps build the automation tool does not escape the talent pipeline problem; it may simply arrive there first.
What This Means for Data Businesses
The more immediately investable consequence may sit with the data providers. Rogo and Hebbia now face pressure from both directions, as banks continue building more AI tools in-house and the frontier model providers keep moving up the stack into finance workflows. But the larger data platforms, S&P Global, MSCI, Moody’s, FactSet, are in a different position. OpenAI has said it is optimizing MCP-based integrations with frequently used financial providers such as S&P Global and FactSet for immediate use in the product. Being a native data layer inside a dominant AI workflow tool is a distribution model that legacy terminal contracts never offered.
The Long-Term Verdict
The tool showcases OpenAI’s continued push into enterprise offerings as it positions itself for a potential IPO, though timing remains uncertain. As of mid-August, OpenAI’s enterprise revenue run rate was reported to be up 50% quarter to date. Financial services is not a side bet; it is one of the most concentrated pockets of enterprise software spending on the planet, and OpenAI just planted a flag there with two of the industry’s most respected names attached to the launch.
For Morgan Stanley and Evercore, the calculus is a wager that co-designing disruption gives them durable workflow advantages their competitors will spend years catching up to. That may be right. It is also possible they spent considerable institutional capital teaching a very capable model exactly how to commoditize their highest-margin production work. Both outcomes are live. The firms that built the tool will find out first.
