Hard Questions: The New Skill for the Future
The answer is now cheap. The question is the skill. The evidence says chief executives are losing it faster than their teams.
The answer is now cheap. The question is the skill. The evidence says chief executives are losing it faster than their teams.
At the 2026 Forbes Future of Work Summit, Hilton's Chief Human Resources Officer Laura Fuentes made a claim most chief executives would nod along to and then ignore. In the age of AI, knowing how to ask the right question is what sets people apart. Answers are easy to find. Questions are not.
I would go further. The quality of questioning is no longer a talent issue for the people function. It is the chief executive's own operating discipline, and most are losing it.
Here is the mechanism. Every generative AI tool rewards you with a fluent, confident, complete answer in seconds. The fluency arrives before anyone has checked whether the question was right. The answer looks finished. The thinking never started. I call this the illusion of fluency, and it is now the most common failure I see in boardrooms from Riyadh to London.
The cost is already on your P&L.
SurveyMonkey's 2026 State of Curiosity report, 1,925 US workers surveyed in April, found that half had been forced to redo work because the right questions were not asked at the start. Forty-six per cent had watched time or money wasted because an assumption went unchallenged. That is rework and write-offs, not culture.
The same report exposes the senior shortcut. Directors and vice presidents are nearly three times as likely as individual contributors to ask an AI tool instead of a colleague: 33 per cent against 12. More than a third of AI users then accept the output as it stands, or after a quick check. The people with the most authority are the people most likely to bypass a human, least likely to interrogate the answer, and most able to act on it at scale.
The silence runs downward too. Forty-four per cent of workers believe asking too many questions makes them look incompetent. Among Gen Z, 41 per cent admit to pretending they understand something they do not. A workforce that will not ask, led by executives who no longer need to, is an organisation flying on assumptions.
Microsoft and Carnegie Mellon added the cognitive layer in February 2025. Across 319 knowledge workers and 936 real tasks, the more confident people were in generative AI, the less critical thinking they applied. Confidence in the tool displaced scrutiny of the result. That is atrophy, and it compounds.
None of this is new. Francesca Gino's 2018 study of more than 3,000 employees found only 24 per cent felt curious at work on a regular basis. What is new is the tool that makes the barrier invisible. Nobody has to admit they did not ask. The machine answered.
Most companies will file this under learning and development. That is the mistake.
A chief executive makes perhaps a dozen decisions a year that move the enterprise. Each is now AI-assisted somewhere in its chain: the market analysis, the due diligence summary, the board paper, the scenario model. If the question at the top of that chain was wrong, every fluent answer beneath it is precisely wrong. Speed delivers the wrong answer sooner and with more confidence.
Professor Megan Reitz of Saïd Business School, Oxford, wrote a line about Hidden Power that I have not been able to shake: every organisation is one unasked question away from its biggest mistake. When the chief executive stops asking, nobody below will start.
Chief executives are also the one group whose questions are never refused. A junior analyst who asks a pointed question risks looking incompetent to 44 per cent of the room. The chief executive's questions are guaranteed a hearing. That makes them the cheapest, highest-return intervention available.
And boards now ask how AI-assisted decisions are made. The honest answer in most companies is that nobody examined the prompt. Questioning is no longer a personality trait. It is a control.
Six things sit underneath a hard question, and only two are about the question itself. Specifying the problem in one sentence before anyone opens a tool, because most hard questions fail there. Treating a confident, complete answer as a hypothesis, a habit the Microsoft data shows weakens as trust in the tool rises. Knowing the balance sheet or the supply chain well enough to see where the answer is thin. Being able to say “I do not understand this yet” in a room where 45 per cent of younger workers feel pressure to already know. Noticing what nobody has raised and which assumption the whole discussion rests on. And timing: one question, cleanly, while it can still change the decision. A question after the decision is a complaint.
The first two can be taught in weeks. The fourth and fifth are courage and attention, and the chief executive has to model them, because nobody below will go first.
None of these need a programme. They need three habits, done visibly.
The test for all three is the same. In the last month, which decision did you reverse because someone asked a better question than you had? If you cannot name one, the machine is already doing your thinking, and your organisation can tell.
For two decades the premium went to the executive with the answer. The answer is now a commodity your competitors also own. What is scarce, and getting scarcer, is the executive who can specify the problem, doubt the fluent reply and ask the question nobody else in the room would risk.
Asking hard questions is not a soft skill. It is the new hard one, and it is yours to model first.
For Board Advisory, Strategic Consulting, Keynote Speaking or Private Executive Advisory enquiries, please write with a short note on the decision, audience or transformation context. Suitable engagements are shaped through an initial private conversation.