The analyst layer in investment banking is where a deal's supporting infrastructure gets built: research compiled, financial models constructed, pitchbooks assembled before any client conversation begins. It is also where junior banker hours concentrate, and where the cost structure of a deal team is most exposed to automation. OpenAI has launched ChatGPT for Financial Services to compete for that workload.
The product targets the research, modeling, and pitchbook tasks that banks have historically staffed with junior analysts. These deliverables sit toward the lower end of the deal hierarchy in terms of judgment required, but near the top in hours consumed. A language model that can compress that cycle time carries a clear cost argument to a bank's technology buyer.
Where the model enters the stack
Pitchbook assembly and financial research synthesis follow repeatable patterns. The output is formatted and structured, but the inputs vary enough with each deal that banks have not found a way to template the labor out of the process. That structural gap is the position OpenAI is occupying.
Whether ChatGPT for Financial Services can meet the precision requirements that financial modeling demands is the question the product now has to answer in live deployments. A model that produces fluent but numerically imprecise outputs does not actually compress analyst hours. It relocates the burden to verification, which may cost as much as it saves.
What OpenAI describes as its target is the junior end explicitly: research aggregation, model construction, and the pitch decks that precede senior banker involvement.