There’s a reasonable argument that AI should narrow the gap between average and exceptional people. Give everyone access to the same extraordinarily capable researcher, analyst, writer and coder, and surely everyone starts looking a bit more talented.
I’m beginning to think the opposite might happen. AI can give almost anyone considerably more capability, but giving someone access to extraordinary intelligence doesn’t automatically make them more intelligent. Occasionally it just allows them to produce a poorly thought-through answer considerably faster.
I’m seeing this first-hand in executive interviews. Almost every candidate I meet now uses AI. What’s striking isn’t who uses it, but how they use it. Most jump remarkably quickly from problem to prompt, without spending much time establishing the ground rules first.
What exactly are we trying to achieve? What context does the machine need? What assumptions are we making? What are the constraints? What does a good answer look like? And how are we going to know whether what comes back is actually any good?
The better candidates tend to establish those things first. They frame the problem, provide the context, challenge assumptions, interrogate what comes back and apply their own judgement before deciding what to do with it. The technology may be identical. The quality of thinking wrapped around it isn’t.
What the research is finding
Recent research suggests they’re onto something. Harvard Business Review reported on a study of 2,500 KPMG employees over eight months that found sophisticated AI users weren’t simply using it more. They were more ambitious in what they asked of it, gave it clearer objectives and treated AI as a reasoning partner rather than simply a productivity tool.
Other research has found AI can help people tackle unfamiliar tasks more effectively without magically turning novices into experts. Relevant experience still matters when it comes to interpreting, refining and judging what the machine produces. That distinction becomes more important as AI gets better, because producing something that looks impressive is rapidly becoming the easy bit. Knowing whether it deserves to survive contact with the real world is rather more valuable.
For years, expertise was partly valuable because experts knew things other people didn’t. Knowledge itself becomes less scarce when everyone has a machine capable of knowing quite a lot. The premium increasingly sits with what someone does with it: judgement, critical thinking, curiosity, adaptability, emotional intelligence, strategic rigour and executional discipline.
The premium increasingly sits with what someone does with it: judgement, critical thinking, curiosity, adaptability, emotional intelligence, strategic rigour and executional discipline.
Same machine, very different outcome
None of those are shiny new AI skills. They’ve always separated exceptional people from everyone else. What’s changing is the leverage sitting behind them.
Give two executives exactly the same AI and both can analyse more information, model more possibilities and explore more arguments. One can frame the wrong problem beautifully and receive an extraordinarily comprehensive answer to a question nobody should have been asking. The other identifies the decision that matters, supplies the right context, challenges the assumptions and uses AI to expand their thinking rather than replace it.
Same machine. Very different outcome.
Same machine. Very different outcome.
What this means for hiring
That has implications for hiring. Experience and track record still matter, obviously, but increasingly I’d want to understand how somebody thinks, how quickly they learn, how they make decisions when the answer isn’t obvious and how effectively they can direct the intelligence now available to them.
It’s one of the reasons we’ve spent years assessing critical thinking, emotional intelligence, strategic rigour and executional discipline at High Performance People. Those qualities mattered before AI arrived. But when one person can suddenly access vastly more information and execution capacity, the quality of the judgement directing it becomes considerably more consequential.
There’s another side to this that companies need to own. How effectively someone uses AI isn’t simply a reflection of the individual; it’s increasingly a reflection of the environment the company has built around them. If every employee starts with a blank chat window and has to reconstruct the company’s context, standards and decision rules themselves, you’re effectively asking everyone to reinvent the wheel, albeit with a very clever wheel-building assistant.
That’s one reason we’re building Context Engines around businesses rather than simply giving people access to more AI tools. The company benefits because its knowledge and context become reusable, but so do its people. Every interaction can make the system more useful, while every employee gets to start further up the learning curve. Great people still need to establish the ground rules for good thinking; great companies can increasingly establish some of those ground rules for everyone.
The hiring paradox
Which creates an interesting hiring paradox. Companies may need fewer people to produce the same amount of work, but that doesn’t make talent less important. It makes every person you choose to put into the organisation more important.
If a ten-person team can increasingly produce what once required twenty, getting those ten people right matters rather a lot. Put more intelligence and automation around exceptional people and you don’t just increase their productivity. You increase the amount of value their judgement can influence.
Unfortunately, the reverse works too. AI is wonderfully democratic about giving poor judgement more horsepower.
AI is wonderfully democratic about giving poor judgement more horsepower.
Raising the floor and the ceiling
So businesses may spend the next few years doing two things that initially sound contradictory: reducing the amount of human labour required for some work while competing harder for exceptional human capability. AI may raise the floor by helping more people produce competent work, while raising the ceiling for people who know how to use it exceptionally well.
AI is making knowledge and execution more abundant. The ability to frame the right problem, establish the right ground rules and exercise good judgement remains rather less abundant.
Making everyone faster isn’t the same thing as making everyone better.
Sources
- What the Best AI Users Do Differently—and How to Level Up All of Your Employees — Harvard Business Review (2026). https://hbr.org/2026/03/what-the-best-ai-users-do-differently-and-how-to-level-up-all-of-your-employees
- Gen AI Won’t Make Your Employees Experts — Harvard Business Review (2026). https://hbr.org/2026/03/gen-ai-wont-make-your-employees-experts