Wall Street is offering $25,000 daily for AI trainers who previously worked in the industry.

      Felipe Sinisterra and Dave Wang, both former bankers, are fully booked for the next two months, instructing financial institutions on how to effectively utilize the AI tools they have already acquired. According to a Bloomberg article published this week, the pair is charging banks and investment funds up to $25,000 per day to teach senior staff how to operate AI tools that their organizations have already invested in.

      Their client list to date includes notable firms such as T. Rowe Price, Citigroup, and Bank of America. The premise of their work is, on the surface, somewhat embarrassing for the wider industry. Over the last two years, major banks have invested billions in AI infrastructure, model licenses, and internal tools, based on the belief that generative AI would transform financial workflows.

      What Sinisterra and Wang appear to be offering is not cutting-edge technology, but rather the knowledge required to effectively use existing systems. In their demonstrations, they teach senior bankers how to utilize commercial models like Anthropic’s Claude, OpenAI’s ChatGPT, and Google’s Gemini for tasks that the bank’s employees have not yet mastered. For example, during one session described by Bloomberg, they analyzed a video pitch from a startup founder using Gemini’s video understanding feature.

      Their credibility stems from their previous careers before consulting. Sinisterra worked at Goldman Sachs and Bank of America before heading fintech investments at SoftBank, where he invested $2 billion and fostered several AI ventures. Wang previously worked at Morgan Stanley and led cryptocurrency initiatives for SoftBank Latin America; he is also on the advisory board of the Harvard Data Science Initiative. Sinisterra operates the training business known as Wall Street Prompt.

      The nature of the demand itself tells a significant story. According to the booking calendar, banks are not struggling to access models; rather, they are finding it difficult to address the practical, ground-level work of integrating probabilistic tools into a profession that traditionally relies on deterministic outputs—an aspect left out of the AI strategy presentations by McKinsey for 2023 and 2024.

      Sinisterra and Wang claim that tasks such as earnings interpretation, market analysis prompts, due diligence synthesis, and pitch deck evaluations are areas where most analyst teams are only utilizing a small fraction of the capabilities of the tools available to them.

      The pricing also serves a specific signaling purpose. A rate of $25,000 per day aligns closely with the fees generated by a single managing director at a large US investment bank over a quarter, suggesting to procurement departments that the expense is negligible and not worth negotiating.

      Additionally, this rate exceeds what the big-four consulting firms charge for similar training sessions, which reflects a broader trend towards smaller, more agile consultancies composed of ex-practitioners taking on AI-related projects outside the traditional McKinsey-Bain-BCG framework.

      The long-term question remains whether the category of trainers will diminish. Anthropic has already begun expanding into financial services since early 2026, forming partnerships with Moody’s and achieving complete integration with Microsoft 365. As model vendors move towards providing plug-and-play financial workflows, the value of customized prompting tutorials is expected to decrease.

      So far, Sinisterra and Wang have maintained their advantage by highlighting live, unique use cases that go beyond vendor documentation. How long this knowledge gap remains is uncertain.

      For now, banks are essentially paying for the two-month waitlist. The actual training, as highlighted in several recent client testimonials, could be duplicated by a moderately curious analyst with access to a corporate ChatGPT license and a weekend to spare.