Driving AI adoption from 40% to 85%

A 20,000-person Fortune 500 software company. Dozens of AI tools deployed. After months of comms, adoption hovered around 40%.

"We need to upskill the company on AI and become AI-first."

We dug deeper and employees told us the barriers to AI adoption in their own words: I don't have the time. I don't know how. It's too complicated. I don't know where to start. It feels like cheating. Only some of us use it. It doesn't fit my job. The gap wasn't really a comms or knowledge problem. It was a context problem, a permission problem, a "show me, don't tell me" problem. People were waiting to see what AI use looked like in their job, with their peers. Standard upskilling was the wrong move. We also heard from high adopters that they learned AI from their peers — which was key to designing the interventions. And we defined what adoption actually means and how to measure it.

A peer-led adoption system that ran for a full year, embedded in the flow of work. A company town hall that demo'd employee use cases. Gamified promptathons. AI expos. Dedicated time to experiment. Lightweight practices embedded in meetings people were already in. Adoption went from 40% to 85% in under a year. Sellers who became high adopters spent an extra hour a week with customers — leading to 2.1× contract value. And an estimated 600,000 hours in saved productivity, while maintaining engagement, performance, and positive sentiment related to AI.