Industry3h ago
Bankers Find No AI Outputs Client-Ready
Zero AI outputs from models like GPT-5.4 and Claude Opus 4.6 passed a quality review by 500 investment bankers. The professionals cited imprecise data and factual errors as primary failures. Despite this, over half will use these tools for initial drafts. This confirms a persistent gap between LLM capability and high-stakes financial precision.