The human moat isn’t built by accident

Why firms now have to teach, on purpose, the judgement that routine work used to teach by default.

In short: AI has removed the routine work through which junior professionals quietly learned to think. The qualities that now define professional value (judgement, trust, ownership and original insight) were a by-product of that work, and nothing has yet replaced it. Firms that want those qualities in their future leaders will have to teach them deliberately, at scale and with rigour. The accident that used to do it for them is over.

The first half of this story is well understood. Checking contracts, building models, first-pass research, formatting the deck: a machine can now do all of it faster and more cheaply than a junior. Most firms are already redesigning their delivery models around that fact.

What has had far less attention is what that work was actually doing for us. AI hasn’t just automated the routine. It has made visible something professional services firms have avoided examining for decades, and it raises a question every managing partner will eventually have to answer: if the task has gone, where will the next generation’s judgement come from?

My answer is that it will have to be taught. Let me explain why, in three steps.

Professional value now sits in the “human moat”

An idea has been gaining ground over the past year or so: the “human moat”. It describes the advantage that remains once analytical work becomes a commodity, and I think it is broadly right. Four things sit at its core.

The first is judgement under ambiguity, where the data is incomplete and the context goes unspoken. The second is relational trust, built through reputation and presence. The third is ownership, the willingness to sign off on a decision and carry consequences that an algorithm never will. The fourth is original insight, the first-hand experience that can’t be recycled from somebody else’s summary.

None of it scales. None of it automates. All of it compounds.

The strategic logic is sound. The value of a professional is shifting from producing the answer to being accountable for it. The difficulty is that nobody is building the moat.

The mechanism that used to build it is being switched off

For decades, junior professionals learned their craft by doing routine tasks, and everyone assumed the repetitive work itself was the training. It wasn’t. The task didn’t build good judgement. The thinking the task forced on people did: asking why, spotting the risk, comparing against precedent.

I suspect most senior people reading this learned their judgement that way. Not from the spreadsheet, but from the manager who looked at it and asked what it actually meant. The task was only ever the vehicle, and AI has now taken the vehicle away.

That leaves firms with a quiet but serious gap. The work that once produced judgement as a by-product is disappearing, and nothing has taken its place. So, we find ourselves in the slightly odd position of celebrating the four pillars of the human moat as the future of professional value, while switching off the very mechanism that developed them.

Firms that treat this as an efficiency exercise will, I think, optimise themselves into a leadership vacuum. In five to ten years they will find they have too few people who can carry the weight of a decision, for the simple reason that those people never had the chance to learn how.

So the thinking has to be taught on purpose: deliberately, at scale and with rigour

The question for every firm, then, is not “how do we use AI?” That one is already being answered. The more important question is this: how do we deliberately teach the thinking that the task used to teach by accident?

Whatever the answer, it has to meet three tests. It has to be deliberate, because the accident is over. It has to be scalable, because a small number of experienced mentors cannot coach every junior through every problem. And it has to be rigorous, because judgement is not a personality trait. My background is in psychology, and I would put the answer firmly in the category of learned disciplines: built through structured analytical thinking and problem-solving, and communicated through a clear story that earns trust.

That is the problem our Analytical Problem-Solving (APS) eLearning programme, launching soon, is designed to solve. For over 35 years we have taught the analytical thinking and structured problem-solving that separate high performers from everyone else, using the same methodology the world’s leading consulting firms rely on. APS makes that capability available to every professional in your firm.

Each learner works through a real-life case study with an agentic AI coach alongside them, guiding their learning and pushing them to think more rigorously. The coaching style is the point. The coach doesn’t hand over the answer. It asks the questions a good senior would ask: why does this matter, what is the risk, what does the evidence really say, what would you recommend and why? That is exactly the thinking the routine task once forced on juniors. Now it is taught on purpose.

The irony isn’t lost on us. We are using AI to build the very skills that AI can’t replace.

So what?

The firms that tackle this directly will be the ones with leaders ready to step up when others find the bench empty. The firms that treat it as a design challenge rather than a cost saving still have time to act. The window is open, but it won’t stay that way.

The human moat won’t build itself. Someone has to teach it.