62% of Finance Workers Say AI Errors Reached Clients or Decision Makers

Macabacus said in an announcement that 62% of surveyed finance workers believe an AI caused error reached a client or internal decision maker during the past 12 months. Another 46% believe an error probably entered a final product without being detected.

The report found that 87% use AI daily or weekly to create financial models and client presentations. Only 23% reported comprehensive safeguards covering approved tools, accuracy checks, brand compliance and review workflows.

Among analysts and associates, 43% said AI increased confidence in their work. That figure fell to 29% among vice presidents, directors and managing directors, a group also more likely to say AI had weakened their confidence.

Macabacus issued the release in German, and the linked source is that edition. Two further things about it belong with the numbers. The survey population is Macabacus's own base of 75,000 users, not the industry, so it describes people who already buy financial productivity software. And Macabacus sells the review and verification tooling that the report finds firms lack, so the finding runs toward its own product. What the figures do capture, whoever collected them, is a gap between how many people use these tools daily and how many work under any checking process, and both halves come from the same respondents.