Using OpenAI o1 for financial analysis

Wall Street has spent two years throwing chatbots at spreadsheets with mixed results. Rogo, a startup built specifically for financial analysts, thinks it found a better foundation in OpenAI's o1 model, and the case study OpenAI published makes a fairly narrow but telling argument: reasoning models are better suited to finance work than the faster, cheaper chat models everyone defaulted to first.

The reason is not exotic. Financial analysis is not really a writing task, it's a multi-step logic problem dressed up as one. Pulling numbers from a 10-K, checking them against a company's own guidance, then building a valuation model requires the kind of sequential reasoning that trips up models optimized purely for fluent text. Rogo's product apparently leans on o1 for exactly those moments, when an analyst needs the model to actually work through a problem rather than pattern-match a plausible-sounding answer.

That distinction matters more in finance than almost anywhere else, because a fluent wrong answer about a discounted cash flow assumption doesn't just look bad, it costs money or gets someone fired. So Rogo's bet is that accuracy on multi-step tasks, even at higher latency and cost, beats speed. That's a rejection of the assumption that AI tools need to feel instant to be useful, and it's a signal about where reasoning-focused models actually earn their premium: not chat, but work.

OpenAI, for its part, gets a marquee example of o1 doing something other than winning math benchmarks. Enterprise buyers in finance are notoriously conservative and slow to adopt anything that touches real money, so a working deployment inside investment research is a credible signal to the rest of Wall Street that these models can handle domain-specific rigor, not just general chit-chat.