We measure the distance between human work and machine capability.

Every company is a collection of decisions. Thousands of people, every day, reading context, weighing tradeoffs, choosing what to do next.

AI is learning to make those same decisions — faster each quarter, across more domains than anyone predicted. But no one is keeping score.

We go inside companies, decompose work into its atoms, and build evaluation systems that track AI against real human performance. Not on benchmarks. On your actual work.