OpenAI launches ChatGPT for Financial Services, targeting Wall Street research workflows

OpenAI rolled out ChatGPT for Financial Services, a version of ChatGPT tailored to banking and buy-side workflows that pulls from integrated data sources, cites its evidence, and drafts pitch materials.

The product is aimed squarely at the research and pitch-book layer of investment banking and asset management, work that today sits inside FactSet, Capital IQ, Bloomberg, and a stack of internal templates.

OpenAI's framing puts provenance at the center. The example the company uses is a banker running a P&L normalization: the analyst needs to see the reconciliation behind adjusted EBITDA, identify which costs management excluded, and decide how those adjustments should carry into valuation. The goal, per OpenAI, is not to hand back a single EBITDA figure but to preserve enough source-level detail that a professional can inspect the evidence and form their own view.

That framing matters because it is the framing incumbents have used to defend their moats. FactSet, S&P, and Bloomberg have long argued that the analyst's job is judgment applied to auditable data, and that any AI layer over financial workflows lives or dies on citation, permissioning, and data lineage.

On the benchmark front, OpenAI disclosed results on OfficeQA Pro, which it describes as an independently developed evaluation of 133 questions covering roughly 89,000 pages of U.S. Treasury Bulletins spanning nearly a century. GPT-6 Astra scored 69.9% correctness in OpenAI's evaluation, ahead of GPT-5.6 Sol at 60.2% and Claude Fable 5.1 at 62.4%, gaps of 9.7 and 7.5 percentage points. Benchmarks self-reported by the model vendor should be read with the usual caveats; buy-side diligence teams will run their own.

The launch lands into a crowded field. FactSet introduced FactSet AI for Banking with Finster AI in March as a secure environment for automating investment-banking and research workflows. In July, S&P Global Market Intelligence announced a partnership with Farsight to generate pitch decks, confidential information memoranda, and valuation materials from within Capital IQ Pro. Bloomberg has its own model track. The vertical-AI-for-finance category is no longer contested at the edges, it is contested at the core.

What to watch: enterprise deployment announcements from named banks and asset managers, benchmark responses from Anthropic and Google, and whether FactSet and S&P Global accelerate their AI product cadence in the next earnings cycle.