Banks are expanding AI use in junior workflows, but benchmark data still show limits in client-ready investment-banking output.
Key Highlights
- Deloitte estimated generative AI could lift front-office productivity at top investment banks by 27% to 35%.
- BankerToolBench tested nine frontier models on junior investment-banking workflows.
- The benchmark was built with input from 502 current and former investment bankers.
- Human reviewers rated 0% of tested AI outputs as client-ready.
Wall Street’s AI hiring debate is shifting from whether models can automate junior tasks to whether they can produce work that banks can safely show clients.
Deloitte estimated that generative AI could increase front-office productivity at top global investment banks by 27% to 35% by 2026, highlighting the size of the potential efficiency gain.
New benchmark data show why banks remain cautious. BankerToolBench evaluated nine frontier models on junior investment-banking workflows, with tasks developed using input from 502 current and former investment bankers.
The results exposed a reliability gap. The study found that even the strongest model failed nearly half of the rubric criteria, while bankers rated 0% of tested outputs as client-ready.
That creates a hiring problem rather than a clean replacement story. AI can speed up document drafting, initial analysis, data gathering and formatting, but investment banking also requires judgment, client context and accountability.
Large banks are still increasing AI deployment. Recent banking commentary showed AI is shortening technology project timelines and helping automate financial analysis, but institutions remain cautious about over-reliance and rising model costs.
The practical outcome is likely to be a redesign of junior roles. Analysts may spend less time assembling first drafts and more time verifying numbers, checking source material and controlling final output quality.





_11_13_2026_20_30_45_840725.jpg)
Please wait processing your request...