AI FAILS INVESTMENT BANKING TEST, NONE CLIENT-READY

The evaluation assessed how leading AI systems handle routine junior banker tasks. Results were consistently subpar, with outputs containing imprecision or factual errors that made them unsuitable for direct client use. The findings highlight a gap between AI capabilities and financial services requirements, where accuracy and reliability are paramount. Clients expect error-free analysis and recommendations. However, the test revealed partial utility. More than 50% of bankers indicated willingness to leverage AI outputs as drafts or initial frameworks, requiring human review and refinement before delivery. This middle ground suggests AI's near-term role in banking: augmenting analyst workflows rather than replacing junior-level work entirely. The benchmark demonstrates current models cannot yet fully automate entry-level investment banking tasks but may reduce manual labor when paired with human oversight.