Most organizations sit on rung two of the AI adoption ladder and believe they are on rung three. The gap matters because the interventions that work at each rung differ, and the most reliable diagnostic is what happens when AI fails.