The data showed a pattern. It did not explain the people inside it.

Essay

A dashboard shows that engagement has fallen. Collaboration data shows that a team is becoming quieter. Exit interviews reveal a pattern after several people have already left.

All of that information may be useful. None of it can tell a leader, on its own, what the experience meant to the people inside it.

Data can tell you where to look. It cannot have the conversation for you.

A pattern is not an explanation

AI can process large amounts of behavioural and organisational data. It can surface changes that a leader might otherwise miss. The problem begins when the signal is treated as an explanation.

A lower score may reflect frustration, fatigue, mistrust, a change in the work or something unrelated to the programme being measured. A summary can compress those possibilities into a clean account and still be wrong about the reason.

The more reporting layers information passes through, the easier it is for uncertainty and disagreement to disappear. A polished summary can leave a leader feeling informed while removing the details that would have prompted a real question.

What I mean by strategic empathy

Strategic empathy is the discipline of understanding how a decision is likely to be experienced before treating that understanding as an input to the decision.

It is not agreement, emotional performance or an instruction to avoid difficult choices. A leader may understand the effect of a decision and still conclude that it is necessary. The difference is that the consequences, context and knowledge held by the people affected are considered before the decision is settled.

This matters because people close to the work often know where the process breaks, which exception has been ignored or why a seemingly sensible change will create a new problem. A score may point to the issue. It cannot supply that local knowledge.

Use the signal to choose the conversation

When the data shows a change, resist the urge to explain it immediately. Start with a narrower set of questions:

• What exactly changed, and when?

• Who is closest to the work behind the pattern?

• What might the measure be missing?

• What decision would change if our first interpretation proved wrong?

Then speak with the people who can answer. The purpose is not to collect a few quotations after the analysis is complete. It is to test the analysis before leadership acts on it.

Human context should change the decision

Listening becomes performative when the decision is already closed and the only open question is how to communicate it. If leaders ask for context, they should be clear about what can still change and what cannot.

That does not give every person a veto. It gives leadership better information about consequences, implementation and the assumptions built into its own view.

It also creates a more honest explanation. A leader can say what was heard, what changed as a result and which trade-offs remain. People may still disagree, but they can see how the decision was made.

Know what AI is removing

AI can help organise interview notes, compare themes or identify questions worth pursuing. It can also smooth away the hesitation, contradiction and unusual detail that signal where deeper understanding is needed.

The leadership question is therefore not simply whether to use AI in the process. It is which parts of the human account must remain visible, who will examine the source material and where a summary is not enough.

Do not ask the measurement to do the work of the relationship.