There have been many recent stories about executive leaders mandating AI use. Recently, I saw a report that 73% of executives report underwhelming ROI from AI. Does that mean that companies are wrong to push for AI use? My answer is: “No, but many companies are not doing the push correctly, which is why the results are often underwhelming.”

If your plan is to tell people they must adopt AI, your strategy will fail. You need a nuanced approach that is flexible enough to adapt as you roll out AI. For as Field Marshal Helmuth von Moltke the Elder said: “No plan survives first contact with the enemy.”

There is a fear among employees, stoked by news of layoffs, that AI adoption may cost them their job. Those concerns are not going to help drive enthusiastic adoption. Issuing mandates wrapped in threats that those who do not adopt will be shown the door merely increases those fears. People whose jobs depend on meeting some AI usage metric will find ways to meet those numbers, but that is very different from using AI effectively.

An effective AI adoption strategy does start with a statement from leadership, but that alone is not sufficient; it is not even one of the most important parts of the plan. Communication is key to driving adoption.

Managers need to encourage their staff to share their success with AI, to create a positive atmosphere around AI usage. Having stories drawn from within the company are more persuasive than sales pitches. The more important part of manager communication is for leadership to listen.

Listening to feedback is vital. When people have issues with AI use, those problems need to be heard and addressed. AI is new and evolving, so your processes need to adapt based on lessons learned. The best way to get lessons learned early is to encourage blameless feedback, listen to that feedback, and act on it. Those changes may include an acceptance that, for now at least, there are parts of the business for which AI is not the best fit. No plan should be “AI everywhere for everything.”

Training is essential. An edict from the top may push people to use AI; it will not cause people to use AI effectively. Ongoing training, adapted for your own company use case, is essential to a successful adoption.

Never evaluate people by their AI use. It was never sensible to say Enginneer A is better than Engineer B because they wrote more lines of code or closed more tickets. It makes no more sense to look at token usage as a meaningful measure of contribution. For a manager, they need to evaluate the overall contribution that person makes and not focus on numbers that are not true indicators of progress, and can be easily gamed.

In conclusion, successful AI adoption demands more than top-down mandates or simplistic metrics. It requires a thoughtful, adaptive strategy grounded in open communication, genuine listening, and blameless feedback loops that evolve with real-world lessons. By celebrating internal successes, addressing legitimate concerns without threats, and evaluating contributions based on overall impact rather than token counts or usage quotas, leaders can foster an environment where AI truly amplifies human potential.

Ultimately, the organizations that thrive will be those that treat AI integration as a collaborative journey, not a compliance exercise, ensuring sustainable results that benefit both the business and its people.