Companies routinely promote people because they are good at their work, then give them responsibility for other people without teaching them how to lead. AI raises the stakes: it demands judgment, delegation, and feedback while potentially removing the early career experiences through which people develop those abilities.
That is the problem I keep coming back to after a career across technology, healthcare, consulting, and other industries. We put considerable effort into developing expertise. We are much less deliberate about developing the ability to help someone else succeed.
A strong developer becomes an engineering manager. A successful analyst becomes a department head. Their accomplishments justify recognition, compensation, and opportunity. But those accomplishments tell us little about whether they can handle conflict, explain a difficult decision, build trust, or help an employee work through a problem.
Those abilities can be learned. A promotion simply does not teach them.
Compensation structures can make this worse. When management offers the clearest path to advancement, people face a difficult choice: keep doing work they enjoy with limited opportunities to progress, or take responsibility for people because that is how they move forward. The organization gets a new manager without necessarily getting someone prepared for, or interested in, the work of leadership.
Even people who want to lead need practice and guidance. Someone has to help them recognize when their intervention is useful and when it prevents an employee from learning. They need opportunities to have difficult conversations, reflect on what happened, and try again with better judgment.
Some of that preparation can begin long before someone becomes a manager. Being led well teaches us something about leading. We experience a clear assignment, a thoughtful question, a correction that helps rather than humiliates. We watch someone explain a tradeoff or take responsibility when a decision goes wrong.
That education has always depended too heavily on luck. AI gives us another reason to make it deliberate.
Consider an assignment that once went to a junior employee: investigate a problem, build an analysis, bring back a recommendation. The finished work mattered, but so did the conversations around it. Why did you choose those assumptions? What did you leave out? What would change your recommendation?
If AI produces the analysis and an experienced employee checks it, the organization may finish faster. But unless someone deliberately involves the junior employee, the learning opportunity can disappear with the assignment.
Meanwhile, using AI well requires several of the same habits we should already be teaching future leaders. Define the outcome. Provide context. Decide what to delegate. Review the result. Give useful feedback. Take responsibility for the decision.
A person who can produce an answer has not necessarily learned how to evaluate one. A person who can complete a task has not necessarily learned how to direct work.
Organizations need to build those abilities into everyday work. Give prospective managers responsibility with an experienced leader available to coach them. Discuss actual decisions and difficult conversations while they are happening. Recognize managers for developing people who can exercise judgment independently.
As AI changes assignments, redesign the learning around them. Ask an employee to assess an AI-generated recommendation, defend an alternative, or explain what evidence would make them reject the result. Let them participate in the reasoning that experienced colleagues might otherwise keep to themselves.
This requires time from people who are already busy. It also requires treating the development of others as part of their job.
We cannot keep assuming that capable employees will somehow become capable leaders. We have to teach them, give them room to practice, and hold their managers responsible for helping them grow. As more work moves to AI, that responsibility becomes more urgent.