How Do You Know Who Your Top Performers Are, When AI Writes Everyone's Work?

When polished outputs are available to everyone, traditional performance signals start to disappear. Leaders need new ways to identify real capability.

I recently facilitated a leadership workshop on people management in the AI era. One question kept coming up, in different forms, from different managers:

With AI tools available, how do you know who your highest performers actually are?

A couple of years ago, this was a much more straightforward question to answer.

How we used to judge performance

Think about what performance evaluation quietly relied on: hours of work, the analysis, the thought process, the drafts, the struggle, the employee questions along the way.

You could see someone thinking. The rough first version. The questions they asked. The three revisions before the final document. The effort was visible, and the quality of the effort told you something about the quality of the person.

Today, someone can deliver a polished report, presentation, recommendation, or proposal in a fraction of the time.

And most times, the output looks excellent.

The two questions every leader now faces

Question 1: Is it indeed excellent?

Looking polished and being excellent are no longer the same thing. AI is exceptionally good at producing work that has the shape of quality — confident structure, professional language, plausible conclusions. Whether the substance underneath holds up is a separate question, and answering it requires real expertise from the evaluator.

Question 2: What exactly are you evaluating?

The final document? The thinking behind it? The judgment? The verification? The decisions made along the way? The prompt?

This is the question I see making leaders genuinely uncomfortable — and it should. Because the honest answer for most organisations is: we never had to decide before. The document and the thinking used to be inseparable. Now they can be produced by different parties, one of whom is a language model.

The uncomfortable reality

One thing that surprised me during the workshop was how quickly managers stopped talking about AI and started talking about trust. 

Which raises questions most organisations haven't confronted yet.

Could some of the people currently rated as top performers simply be the best at producing impressive-looking outputs?

And could some of the strongest thinkers in the organisation be getting overlooked — because their value isn't as visible in an AI-polished world?

If your performance system rewards deliverables, and deliverables are now cheap to produce, then your performance system is measuring something that has quietly stopped mattering.

What can leaders actually do?

I don't think the answer is banning AI tools or demanding to see prompts — that fight is already lost, and it misses the point. The tools are genuinely valuable. The problem isn't that people use them; it's that our evaluation habits haven't caught up.

A few directions that came out of the workshop discussion:

Evaluate the conversation, not just the artifact. Ask people to walk you through their reasoning. Why this recommendation and not the alternative? What did they reject, and why? Strong thinkers can defend their work in depth. Output-producers struggle the moment you leave the document.

Watch what happens when things go wrong. Judgment reveals itself in ambiguity — the messy client situation, the analysis that contradicts expectations, the moment the polished plan meets reality. AI helps far less there, and real capability shows.

Value verification as a skill. In an AI-assisted workplace, the person who catches the confident-but-wrong number is worth more than the person who produced ten flawless-looking reports. Make checking, questioning, and validating visible parts of the job — not invisible overhead.

The leadership challenge isn't getting smaller

It's becoming harder to see.

So let me leave you with the question I asked the workshop, and the one I'd ask any leader reading this:

Which person in your team are you most confident is a high performer — and what evidence are you basing that confidence on?

If your honest answer is "their outputs look great," it might be time to look closer.

I run leadership and AI-readiness workshops for finance teams and organisations through Forte Consultum, covering exactly these questions — how to manage, evaluate, and develop people in an AI-augmented workplace. If this is a conversation your leadership team needs to have, get in touch.