MIT and Stanford researchers published a case study of generative AI's impact on 5,179 customer support agents at an unidentified Fortune 500 software company, finding a 13.8% gain in problems resolved per hour when AI provided real-time guidance prompts. The effect was concentrated in less-experienced and less-skilled workers: agents with two months of experience matched six-month veterans when using AI. Turnover fell, particularly among newer staff, and customer escalation rates declined. High-skilled agents saw minimal direct productivity lift. This represents one of the first empirical measurements of AI's labor-market productivity impact in a real operating environment rather than laboratory or task-based settings.
258 other tracked developments touching the same claims.