AI won't fix your mindset

TLDR — Plenty of organisations are reaching for AI the way they once reached for a branded methodology: as something to buy so they don't have to do the harder thinking. But AI makes the "how" almost free, which means the "how" stops being your advantage. What's left is the part no tool can hand to you: the clarity to know what you're trying to do, the pace to keep moving while you find out, and the human trust, credibility and judgement that make either of those deliver meaningful value.

In every board room right now, there is a discussion about their "AI strategy." Not because anyone in the room can clearly say what it's for, but often because everybody else seems to have one, that the media or big consulting firm have told them they need one and that not having one feels dangerous. So a budget appears. A head of AI is appointed. A pilot is commissioned. Activity rises, confidence spikes, and a year later not very much has changed, except that there is now a dashboard nobody reads and a licence bill that the CFO is starting to ask questions about.

If that sounds familiar, it should. We have seen this film before, maybe even starred in an early version. It is exactly what organisations did with any branded methodology: reach for something you can buy, so you don't have to do the harder thinking underneath it. AI is just the newest, shiniest thing to put on the credit card. And the moment a technology gets sold to you as the answer, rather than as a tool, you should be very suspicious, because someone has quietly removed the difficult part and overcharged you for the easy one.

AI makes the 'how' free, so the 'how' stops being the edge

Here is what AI genuinely changes, and it is worth being precise about it.

Most of these tools are trained on a vast pile of yesterday's best practice. That is a real and amazing gift. The "how to do something," the method, the technique, the worked example, is now a prompt away for almost anyone. Things that used to take a specialist a week are now a conversation. That is democratisation, and it is mostly good.

But let's think about the consequence. If the method is now available to everyone, instantly and cheaply, then the method can no longer be the thing that sets you apart. The advantage doesn't disappear, it just moves. It moves to the two things AI can't hand you across the table.

The first is clarity. AI will cheerfully give you a brilliant answer to a badly chosen question. It has no idea whether you are pointed at the right thing, because that judgement, the "so what are we actually trying to achieve here," is yours and only yours. Give a confused organisation a more powerful tool and you mostly get confusion, faster.

The second is the one almost everyone forgets, and it is the more interesting half. AI cannot move for you. It can draft the plan in seconds, but it cannot make you start, cannot give you the nerve to ship the imperfect first version, cannot keep you going when the first attempt is wrong. The pace at which you act, learn and course-correct is now a bigger differentiator than the method itself, precisely because the method got cheap and pace did not.

Faster tools, same stuck organisations

You can see it already. Teams handed a shiny licence for the latest assistant or magical model are now six months on making the same decisions just as slowly, only now with better-formatted slides. Individuals signing up for the prompt-engineering certificate, which is a new method to learn in order to avoid the old, harder question of what they want. Whole companies announcing an AI transformation in the same breath, and with the same hopeful vagueness, as they once announced a digital one or some tagline of 'mobile-first', 'digital-first' or 'agentic-first',...just give yourself a break!

"When anyone can generate a confident answer in seconds, the question is no longer what's the answer. It's who do you trust?"

And now there is hard evidence for the gap, not just anecdote. An MIT study of enterprise AI in 2025, The GenAI Divide, found that around 95% of corporate generative-AI pilots delivered no measurable return, despite something like thirty to forty billion dollars poured into them. Ninety-five per cent. The tools were extraordinary and the results were not, and the report's own explanation was not the technology. It was the organisations around it: most never learned, never adapted, never changed how they actually worked. Which is the whole point. A better tool, dropped into the same stuck mindset and the same standstill, just gets you a more expensive standstill.

There is even a phrase doing the rounds, "AI psychosis." Used properly it names something genuinely troubling: people spiralling into delusion through long, agreeable conversations with a chatbot that validates whatever they bring to it, and it deserves to be taken seriously rather than borrowed as a bit of boardroom colour. But the boardroom has its own milder affliction worth naming, so let me be plain about that one instead: leaders who have gone slightly unhinged with conviction about the technology while staying strikingly unclear about the point of it. It is the old pattern in new clothes: an undue focus on the tool, a promise of speed, and then the slow, expensive discovery that the tool was never the thing that was missing.

Both of these are human before they are anything else

Here is the part that ties it all together, and it is the part the breathless AI demos tend to skip. Neither clarity nor pace really comes down to one clever person having a good idea. Both are human and relational, and both stand on something AI hasn't got and can't fake: trust, credibility, hard-won domain expertise, collaboration and leadership.

Teams act on trust, not on tooling.

Let's start with clarity. Knowing what you are actually trying to achieve is judgement, and judgement is bought slowly, with experience. It takes someone who knows the domain in their bones to tell a genuinely good answer from a fluent, confident, completely wrong one, and AI will hand you both with exactly the same conviction. The scarce thing is not the answer. It is the credible person who can tell which is which, and whose judgement the room actually believes.

Then mobilising. Nobody ever moved an organisation because a tool told them to. People move for leaders they trust. Clarity has to be carried, argued and owned by someone with the standing to be followed, or it stays a tidy slide that nobody acts on. Pace is not a switch you flick; it is a team choosing to act, and teams act on trust, not on tooling.

And there is a sharper turn underneath all of this. As AI makes plausible content infinite and almost free, the one thing it cannot manufacture quietly becomes the most valuable thing you have: trust. When anyone can generate a confident-sounding answer in seconds, the question stops being "what is the answer" and goes back to the older, more human one. Who do you trust?

The fix

None of this is an argument against AI. Learn it, use it, all of it. The methods are a search away and most are genuinely useful in their place. Just stop expecting the purchase to do the part that was always going to be hard.

So before the AI strategy, do the unglamorous thing. Get ruthless about what you are actually trying to achieve, and be honest about whether you have the clarity to say it in a sentence, and the standing to make it stick. That is the work the tool can't do for you, and it is the work that decides whether the tool helps or just accelerates the mess.

Then move. This is the part I would drive the hardest, because it is where AI quietly changes the game in your favour. If trying things is now cheap and fast, the winners will not be the ones with the best method, they will be the ones who move, learn and adjust quickest. Small steps, fast loops, course-correct, repeat. AI lowers the cost of the first step to almost nothing, which means there is no longer any excuse not to take it.

The "how" is finally free..ish (the cost of tokens notwithstanding, of course!). What it can't sell you is the clarity to choose well, the credibility to be believed when you do, and the trust that gets people to move with you. That was always the expensive part, and it always will be. It just isn't for sale, by anyone, in any box, however clever or pretty the box has become.