What the Way You Write Prompts Says About You — Recruitment in the Age of AI

Last week, I attended a training session at the National University of Singapore led by Akina Ho, an executive coach and associate professor at the NUS School of Computing. The topic: how major consulting firms are now incorporating AI interaction tests into their recruitment processes.

What I heard really struck a chord with me. It was surprising and made perfect sense.

AI hides nothing. It reveals everything.

Before AI entered the recruitment process, an interviewer had to infer the quality of a candidate’s reasoning based on what they said. It was imperfect, partial, and susceptible to preparation, rehearsal, and embellishment.

Today, when a candidate interacts with an AI tool in real time, the recruiter observes something fundamentally different: the very structure of their thinking. What they ask the AI. What they accept without question. What they challenge. Where they stop thinking. What they don’t know.

Ms. Ho put it with a precision that leaves no room for ambiguity: “AI does not improve weak thinkers. It exposes them faster.”

The Johari Window, a tool for assessing AI maturity

To understand this phenomenon, Ms. Ho draws on the Johari Window - a conceptual framework developed in the 1950s by Joseph Luft and Harry Ingham to describe self-awareness in interaction with others.

Reminder: the window distinguishes four zones. What you know about yourself and others see as well (Open). What others perceive but you are unaware of regarding yourself (Blind). What you know but conceal (Hidden). And finally, what is unknown to everyone, including yourself (Unknown).

Applied to the AI context in a recruitment setting, this framework takes on a new dimension.

The Blind zone is the most revealing. A seemingly brilliant candidate may respond superficially, accept the AI’s initial answers without questioning them, never test their own hypotheses, and never frame the problem before querying the tool. This behavior reveals cognitive blind spots that no traditional interview could detect.

We are, in fact, taught to identify and address our blind spots in the Chief Data and AI Officer program at NUS, which I have been attending since September.

The Hidden zone also speaks volumes: the candidate who over-delegates to the AI, who fails to add their own value beyond the model’s raw output, betrays a relationship of dependency that they would likely have concealed in a traditional interview setting.

What Top Firms Really Look For

Firms that use these tests aren’t trying to find out if a candidate uses AI. They’re trying to understand how the candidate thinks with it.

Three key areas are examined. First, the ability to articulate: can the candidate clearly define a problem? A vague prompt yields a vague response, and this is immediately apparent.

 Next, the ability to challenge: does the candidate question the AI’s outputs, compare them to their own judgment, and identify their limitations? Finally, the ability to synthesize: can they transform an AI response into a clear, structured, and defensible executive message?

This is precisely what the Top-and-Bottom Prompting taught in the module seeks to develop: constructing a prompt that combines the role (who is thinking?), the question (what to solve?), and the output format (how to deliver?). This isn’t just a technique. It’s applied intellectual clarity.

What this means for you, right now

If you’re a leader, you should ask yourself the following question: in your upcoming hiring rounds, how will you assess your candidates’ AI maturity? Not their knowledge of the tools, but their way of thinking with them.

If you’re a candidate, the real question isn’t whether you know how to use ChatGPT or Claude. It’s whether your interaction with these tools reveals a rigorous thinker or someone who delegates their thinking without a filter.

The Johari Window teaches us that our blind spots exist, whether we like it or not. AI simply makes them visible, faster, and to those who know how to look.

It’s time to intentionally design our relationship with artificial intelligence. Not for appearances’ sake. To think better.