How to keep your intuition sharp in the age of AI (or why I'm getting my hands dirty again)

Intuition isn't some mysterious gift: it's pattern recognition, built over years of experience and nurtured by feedback loops. When the world starts changing faster than our loops, it silently erodes. Here's how I've witnessed my own intuition erode in the face of artificial intelligence, what research says about this phenomenon, and the method I use to sharpen it. Spoiler alert: it involves getting your hands dirty.

The day I stopped trusting my instincts

About two years ago, I noticed something unsettling. When asked for my opinion on the impact of AI on organizations, on our personal lives, on social and geopolitical balances, my answers came less quickly. Less clear. My intuition, the one I've relied on as an executive coach and facilitator for years, wasn't as sharp in this area anymore.

For someone who spent twenty years in technology before dedicating themselves to people, it was a humbling realization. And a valuable one. Because admitting that you no longer understand what's happening is the first step that allows you to start understanding again.

But before I tell you what I did with this observation, we need to agree on what intuition is, because we attribute powers to it that it does not have.

What exactly is intuition?

Intuition is the ability to instantly recognize a pattern encountered before, without conscious reasoning. Psychologist Gary Klein demonstrated this in firefighters and nurses: their "sixth sense" is actually a library of past experiences, accessed in a fraction of a second. Nothing magical about it. Just compressed experience.

But intuition has conditions for validity. In a now-classic article, "Conditions for Intuitive Expertise: A Failure to Disagree" ( American Psychologist , 2009), Daniel Kahneman and Gary Klein, two researchers who started from opposing positions, agreed on two criteria. Intuition is reliable only if, first, the environment is sufficiently regular to offer stable patterns, and second, the person has had prolonged practice with rapid and clear feedback.

Remember these two conditions, because they explain exactly what is happening to us collectively right now.

Why is AI eroding our intuition?

Artificial intelligence challenges both of Kahneman and Klein's conditions simultaneously. First, the regularity of the environment: the rules of the game change at a pace that renders our patterns obsolete before we've even had time to solidify them. What I thought I knew about "what the technology can do" in 2022 was already wrong by 2023. Second, the feedback loops: most managers consume AI through layers of abstraction (demos, articles, ready-made tools) without ever engaging with the material itself. No direct contact, no feedback, no intuition.

And the phenomenon is massive. According to the NETendances 2025 survey by the Digital Transformation Academy at Université Laval, 52% of Quebec internet users had used a generative AI tool by October 2025, compared to 33% a year earlier. Adoption nearly doubling in a single year is precisely the kind of volatile environment where yesterday's intuition can become a trap.

The result, I observe in brilliant leaders, is an erosion of systemic understanding. You sense you're losing the thread, but the role demands you appear to be holding it. That's where it becomes a leadership issue, not a technology one. Because escaping this trap begins with a phrase few leaders dare to say aloud: "I no longer understand what's happening, and I'm going to investigate."

My method: go down to the roots of the tree

I've always learned the same way: by working my way down through all the layers of a system, right to the tiniest roots, and then back up, one layer at a time, to the leaves and the fruit. It's a generalist's method. I'm not trying to become an expert in every layer; I'm trying to trace a continuous line of experimentation from the bottom up, so that the whole thing becomes imprinted on my body. So that it becomes like a muscle.

It started young. When I copied programs from magazines onto my VIC-20, I was unknowingly developing an intimate understanding of the machine. As a teenager, I spent my evenings on BBSs, those pre-Internet electronic bulletin boards where you connected directly to someone else's computer. I felt like an explorer discovering a new continent. My friends and I already had a surprisingly accurate intuition about what the network would become. But I have to be honest: no one, not even the brightest minds I studied with in engineering, some of whom earned doctorates in artificial intelligence, anticipated the current speed. Intuition provides direction, never a timeline.

With each subsequent wave — the internet, Linux, open source software, social media — I repeated the same process: diving in, exploring, and rebuilding my understanding layer by layer. Then, about ten years ago, I shifted this focus to people, coaching, and facilitation ( I've recounted this shift elsewhere ). And since 2023, the arrival of generative AI has rekindled my passion for technology. What strikes me most is the level of energy I feel: the same sense of wonder I felt at twelve years old, gazing at a cybernetic world under construction. When a technology makes you feel twelve again, it's a signal you need to heed.

My current lab: a second brain and AI agents

In practical terms, my current project involves building my own personal knowledge management system (a "second brain") connected to AI agents. I structure my personal and professional data within it, and above all, I dynamically crystallize what defines me as a person and as a professional: my values, my voice, my methods, my preferences. Codified in this way, all of this becomes usable by AI tools.

The result changes the nature of the relationship with the machine. AI ceases to be a replacement and becomes an assisted intelligence: it produces artifacts that are an extension of myself, with my flavor, my texture, my personality. The difference between the two approaches is enormous, and it isn't acquired by reading articles. It's acquired through experimentation, including making mistakes ( my voice dictation tests can attest to that ).

I'll dedicate a future post to the architecture of this system. What matters here is the principle: this laboratory is my way of recreating the two conditions of Kahneman and Klein. A regular testing ground, and rapid, direct feedback loops. My intuition about AI has been sharpening, measurably, since I've been working directly with the material instead of just skimming the surface.

What if your grease wasn't mine?

Please note: I'm not prescribing my method. Stripping down all the technical layers of a system is the approach of a generalist with an engineering background. It's probably not yours, and that's perfectly fine. Everyone learns differently, and the real challenge isn't copying a formula: it's identifying your own strengths to keep your intuition sharp.

This is exactly the kind of question that a coaching or facilitation process addresses effectively. Not because a coach holds the answer, but because clarifying one's development and learning objectives, defining one's learning style, and choosing an initial, suitable testing ground, all of this is difficult to do alone but much faster with support from a team. In my AI integration coaching, I never provide a demo for the client to copy. The client chooses and implements; I coach the choice and the discernment. It's the same approach I use in management consulting: my job is to structure the thinking before it structures the tools.

The bare minimum starting point, accessible to any leader this week: take a real pain point from your daily management life, small enough to be tested in an hour, and tackle it yourself. Not your IT team. You. Then judge the result against your specific context, not against the demo. You've just created your first feedback loop.

Getting your hands dirty to keep your hands on the steering wheel

Whether we celebrate or fear the direction our technological world is taking, one thing seems pragmatically essential: to influence the direction of things, we must be able to understand them, and to understand them, we must interact with them. We can choose to see AI as a dark path for humanity. We can also choose to see it as a driver of evolution that has the potential to make us more human, provided that humans with clear intentions remain in charge.

Staying relevant in a changing world isn't about keeping up with technology. It's about nurturing intuition: deliberately recreating the conditions in which it develops. It involves direct experimentation, rapid feedback, and the humility to admit when you no longer understand. Getting your hands dirty is optional in its form. But it's not optional in principle.

Do you feel that your framework for understanding no longer keeps pace with the world, and you want to clarify where to start again? Let's talk about it .

---

Frequently Asked Questions

Do you need to be technically skilled to keep your intuition relevant when faced with AI?

No. The key isn't technical expertise, it's direct contact with the subject matter and the resulting feedback. A non-technical leader can develop a solid understanding of AI by experimenting with their own use cases, at their own level, rather than by consuming demos and summaries. The specifics of this hands-on approach vary depending on the individual; the principle of direct contact, however, remains constant.

Why does intuition become less reliable when the environment changes?

According to the work of Daniel Kahneman and Gary Klein (2009) , intuition is only reliable in a sufficiently consistent environment and after prolonged practice with clear feedback. When a technology like generative AI changes the rules of the game faster than our learning loops, the patterns on which our intuition relies become outdated. Intuition continues to speak with the same confidence, but based on expired data.

How can a manager begin to experiment with AI?

Choose a real and recurring pain point in your role, small enough to be tested in under an hour. Personally tackle it using a generative AI tool, without delegating. Then, evaluate the result against your actual context (your objectives, your team, your constraints), not against the demo effect. Repeat this process weekly. It's the loop of experimentation and feedback that builds intuition, not the volume of articles read.

Can coaching help develop one's intuition when dealing with AI?

Yes, indirectly but powerfully. A coaching process helps clarify development goals, identify one's own learning style, and choose suitable areas for experimentation. The coach doesn't provide technical answers; they structure discernment, that is, the ability to judge what AI produces against its own context. It is this discernment, nurtured through practice, that becomes intuition.