Why companies struggle to get ROI from AI: they're automating broken processes instead of redesigning them from the ground up.

What happened

Business leaders gathered at Fortune Brainstorm Tech discussed why many companies are failing to see returns on AI investments. The consensus: firms are rushing to deploy AI at scale without doing the foundational work—organizing data, mapping workflows, and rethinking processes in an AI-native way.

Why it matters

When companies simply drop AI into existing inefficient processes, those inefficiencies get amplified rather than solved. Executives compared this to early industrial electrification, when businesses replaced steam engines with electric turbines and found productivity gains elusive. Real returns require re-engineering workflows end-to-end with domain experts, not just layering AI onto the status quo.

What to watch

Wells Fargo reported a 25% increase in new account openings from AI tools, showing that measurable wins are possible. However, some value—like goodwill from more personalized banker-customer relationships—may not show up in immediate revenue but could drive returns over years or decades. Companies should prepare for the coming year or two when AI agents will be capable enough to power work beyond software development.