Leadership, Lab, Crowd: The AI Adoption Framework

_{IN ONE SENTENCE}

To succeed with AI adoption, you need three simultaneous ingredients: a leadership vision, an experimentation lab, and the crowd of employees using AI daily.

_{THE OBSERVATION}

Most companies approach AI adoption one-dimensionally: either top-down (the CEO decrees an AI project) or bottom-up (employees use ChatGPT in secret). Both approaches alone fail. Top-down lacks ground-level reality. Bottom-up lacks strategic direction. Ethan Mollick proposes a model that combines both, with a crucial third element in between.

The observation is brutal: in most organizations, only 20 to 30% of employees actually use the AI tools provided. The rest either don't see the point or use their own tools secretly to prevent their productivity gains from being used against them.

_{WHAT YOU NEED TO UNDERSTAND}

Leadership: set the direction and incentives

Without a clear vision from the CEO or C-suite, employees have no reason to share their discoveries. Mollick identifies at least seven reasons why people hide their AI usage: fear of being replaced, desire to keep their advantage, concern that efficiency gains will lead to layoffs. Leadership must create an environment of trust and incentives. One company cited by Mollick distributes $10,000 cash every week for the best AI use case.

The Lab: turn prompts into products

Between individual discovery and scaled deployment, you need a lab. This is where the best use cases identified by the crowd are tested, turned into agents, and benchmarked. It's not necessarily heavy technical development: it's structured experimentation. Ideally, the 1-2% of employees who emerge as naturally brilliant with AI become the core of this lab.

The Crowd: give everyone access and harvest the signals

Every employee gets access to AI tools and is encouraged to experiment. The goal isn't immediate productivity but discovery: which uses create value, which fail, what surprises emerge. It's the wellspring of innovation; provided you create the right incentives for discoveries to bubble up.

_{WHAT THIS CHANGES FOR YOU}

• Don't choose between top-down and bottom-up: deploy all three levels simultaneously
• Create concrete incentives for employees to share their AI uses rather than hide them
• Identify your natural AI champions (1-2% of staff) and make them the core of your experimentation lab
• Leadership must use AI itself: companies whose CEO actively uses AI transform faster