The Best AI Users in the Room Are Not Who You Think

Here is the assumption quietly poisoning how most organisations think about AI: that the person best equipped to use these tools is the person who understands how they work. It sounds reasonable. It is almost entirely wrong.

The professionals who are getting the most out of AI right now are not software engineers, data scientists, or anyone who can explain what a transformer model does. They are experienced nurses who know exactly what questions to ask about a patient's history. They are seasoned HR managers who can smell a legally risky answer from three paragraphs away. They are creative directors who know in their gut when a piece of writing has no soul. In other words, the best AI users are people who have spent years developing deep, hard-won expertise in a domain that has nothing to do with technology. Their advantage is not technical. It is professional.

We need to talk about why this is true, what it means for how we think about our own careers, and why the tech industry's framing of AI adoption keeps steering us in exactly the wrong direction.

The Dirty Secret of Prompt Quality

Every AI tool you have ever used lives or dies on one thing: the quality of the input it receives. This is not a technical observation. It is a deeply human one.

When a junior marketing assistant asks an AI to "write a campaign brief for a new product launch," they get something generic and passable. When a marketing director with fifteen years of experience asks the same tool to "draft a campaign brief for a premium skincare line targeting women over 45 who have disposable income but extreme distrust of beauty industry claims, and flag any messaging that risks feeling condescending," they get something genuinely useful. The difference between those two prompts is not technical skill. It is domain knowledge, professional judgment, and hard-earned understanding of what actually matters.

Think about how a skilled lawyer uses an AI research tool. They are not impressed when it produces a confident-sounding summary of case law. They are reading it with decades of pattern recognition running in the background, checking for the specific ways legal language can mislead, knowing which jurisdictions matter and which do not. They are, in effect, supervising the AI the way a senior partner supervises a bright but dangerously overconfident junior associate. That supervisory capacity is the skill. And you cannot download it.

This is the thing the tech industry keeps getting backwards. The pitch is usually some version of: "You do not need expertise anymore, the AI handles that." But the evidence points in precisely the opposite direction. The less domain expertise you bring to an AI tool, the less equipped you are to notice when it is wrong, when it is shallow, when it is confidently producing nonsense dressed up as insight.

What This Actually Means for Your Career

If we accept this, then the career advice that follows is almost the inverse of what most people are currently being told.

We are being told to learn the tools. Update your LinkedIn with "AI-proficient." Take the certificate course. And yes, basic familiarity with the tools matters. We are not dismissing that. But the people who are going to be genuinely irreplaceable in an AI-augmented workplace are not the ones who learned the most shortcuts in ChatGPT. They are the ones who went deep in their field, who built the kind of judgment that only comes from real experience, from mistakes, from years of reading the room.

Consider what happened when a major US hospital system began piloting AI tools to help nurses with documentation and care planning. The nurses who got the most out of the technology were not the youngest, most tech-comfortable ones. They were the most experienced clinicians, who could interrogate an AI-generated care suggestion the way they would interrogate a new resident. They knew what a good answer looked like. They knew what a dangerous one looked like. Their years of expertise did not become obsolete. They became the quality filter that made the whole system work safely.

This is the pattern we keep seeing, in law firms, in design studios, in financial planning practices, in newsrooms. Expertise is not being replaced by AI. Expertise is being amplified by it, for the people who have it, and dangerously papered over for the people who do not.

So here is the question worth sitting with today: Are you investing in getting better at the tools, or in getting better at your craft? Because in five years, one of those investments is going to matter a great deal more than the other. And we would put our money on the craft.