Developing judgement in others

Ask senior leaders where their judgement came from and they will usually talk about decisions they had to own, honest feedback from a manager and the occasions when something went wrong and they had to learn from it. Very little of it came from a course.

We tend to call this experience. But experience on its own does not produce good judgement. People can make the same kind of decision for years and simply become more confident in poor habits. Judgement develops when real responsibility is combined with feedback, seeing how things turn out and the opportunity to reflect on what could have been done differently.

When the experience quietly disappears

AI creates a particular risk here. It is not taking over the hardest, most ambiguous decisions. It is least reliable there. What it can produce quickly is the first draft, the standard analysis or an output that already looks finished. People can therefore find themselves reviewing and approving work they have never had to produce.

That task was doing two jobs at once. It produced a result, and it taught someone how to judge that kind of result. The work still gets finished. What quietly stops is the learning that used to come attached to it.

David S. Duncan writes about this in a recent HBR article. He had expected, as many did, that AI would help less experienced people most. Among his less-experienced colleagues, he saw the opposite: they could produce work quickly but often struggled to tell whether it was any good. His concern is what this could mean over time, with many of the next generation of leaders less able to make decisions when situations are novel or ambiguous.

The distinction is how AI is used. Used to test options, challenge assumptions or provide another perspective, it can support someone’s judgement. Used simply to produce the answer, it can remove some of the practice through which judgement develops.

Trust, and the calls leaders take back

AI is not the only way these opportunities disappear. The same thing can happen when senior leaders do not trust their leadership team to make difficult decisions.

Duncan describes a similar problem in escalation rules, where junior people are required to pass ambiguous or high-stakes decisions up the line. These rules protect quality in the moment, but they can also teach people that uncertain decisions should be passed upwards. Over time, the organisation becomes dependent on a smaller group of experienced decision-makers.

I have seen versions of this in coaching. A senior leader loses confidence in their leadership team and moves back into the detail. Decisions are checked, revisited or pulled upwards. Taking a decision back may protect the immediate result, but it also removes the opportunity for other leaders to demonstrate and develop the judgement their senior leader is waiting to see.

Do it often enough and you can end up with a team that delivers well but still looks upwards whenever the answer is not clear.

The shift from expert to leader

This is one of the harder parts of moving from expert to leader. For years, being handed the difficult decisions was the point. It was how you knew your judgement was valued.

At senior level, the task changes. Your own judgement still matters, but your job is now to develop it in other people. That sometimes means allowing someone else to work through a decision you could have made more quickly yourself.

That does not mean handing over consequential decisions and stepping away. Done carelessly, that leaves people carrying risks they are not ready for. Developing judgement requires boundaries and support.

Give people decisions where enough is at stake for the choice to matter, but not so much that a wrong call is unaffordable. Stay close enough to give honest feedback. Ask them to talk through their reasoning, not just the decision they reached. Let them see the outcome.

Duncan points to fields such as medicine and the military, which have long needed to develop judgement without waiting for costly real mistakes. They use case-based learning, simulation, gradually increasing responsibility and structured reflection afterwards.

Many senior leaders developed their own judgement as a by-product of the work they were given. As that work changes, they will need to be much more deliberate about creating the same opportunities for others.

Source: David S. Duncan, How Do Workers Develop Good Judgment in the AI Era?, Harvard Business Review, February 2026.