In a revealing study that underscores the current limitations of AI in high-stakes professional environments, 500 investment bankers evaluated outputs from leading AI models—including GPT-5.4 and Claude Opus 4.6—on tasks typically handled by junior analysts. The results were stark: none of the AI-generated content was deemed suitable for client delivery. Despite the high accuracy of these models in general use cases, the precision and reliability required in financial analysis were simply not met.
The benchmark focused on real-world tasks such as financial modeling, due diligence summaries, and deal structuring. While the AI models excelled at generating content quickly, the outputs were often riddled with inaccuracies or lacked the nuance necessary for professional use. "The AI was fast, but not reliable," said one senior banker. However, more than half of the respondents indicated they would still use the AI-generated material as a starting point for their own work, suggesting that while AI isn't ready to replace human analysts, it could be a useful assistant.
Implications for the Future of Work
This study highlights a growing tension in the financial industry: AI tools are becoming more powerful, but they are not yet trusted for critical decision-making. As firms invest heavily in AI technologies, the gap between what AI can do and what it can be trusted to do remains wide. The findings suggest that while AI may help reduce the time spent on routine tasks, it will likely not fully automate the analytical and judgmental aspects of investment banking for the foreseeable future.
The results also point to a potential middle ground: AI as a co-pilot rather than a replacement. Investment banks may find value in integrating AI tools into their workflows, using them to draft initial reports or identify trends, while human analysts refine and validate the outputs. This approach could boost productivity without compromising the accuracy and trust that clients demand.
In conclusion, while AI continues to evolve rapidly, this benchmark demonstrates that the path to full integration in finance is still a long one. The industry is at a crossroads—balancing the promise of automation with the need for human oversight and accountability.
