Half of Companies Cannot Prove AI Works — The Measurement Crisis in HR

Companies are spending billions on AI — but can they prove it is working? A March 2026 Federal Reserve study found firms reported just 1.8% productivity gains from AI adoption, and when measured against actual revenue, even that number shrinks. Meanwhile, 56% of CEOs say they have gotten nothing from their AI investments, and nearly half of HR executives have no formal measurement system in place. The problem is not that AI does not work. It is that most organizations adopted tools faster than they built the infrastructure to measure results. Traditional productivity metrics were not designed to capture the diffuse, incremental gains AI creates — the 12 minutes saved per recruiter screen, the 30% faster drafting. Without baselines and attribution models, real gains become invisible. And workforce shifts are already happening in this measurement vacuum. CFOs report AI-driven layoffs running nine times higher than publicly disclosed. Workers are using more AI but trusting it less. HR teams are being asked to lead AI transformation while flying blind on whether it is actually delivering. The companies that win will not just deploy AI — they will build frameworks to prove it delivers. This episode breaks down the measurement crisis, what is driving it, and what HR leaders can do right now.