AI Is Removing the Work That Taught People How to Think

One of the things we've been thinking about recently isn't whether AI will replace jobs. That debate has been done to death. What interests us more is what happens to the people coming behind us.

For years, organisations have talked about removing low-value work. Research, first drafts, competitor analysis, pulling together board packs, cleaning data, sitting through meetings taking notes. None of it was particularly glamorous, and if we're honest, most of us couldn't wait until we no longer had to do it.

The irony is that this was also how many of us learnt our craft. When we think back to the early years of our careers, we weren't learning because someone had designed a brilliant development programme. We learnt because we were close enough to the work to watch experienced people make decisions. We'd build the presentation, sit in the meeting, hear the questions we hadn't anticipated and watch how the discussion unfolded. Half the time we weren't even contributing. We were simply observing how experienced leaders thought.

AI is changing that remarkably quickly. Today, a junior employee can produce a first draft in minutes. Market research that once took days can be summarised before your coffee goes cold. Competitor analysis, meeting notes, financial commentary, board papers; almost every task that used to sit at the bottom of the ladder can now be accelerated by AI.

From a productivity perspective, that's fantastic. I don't think anyone is arguing we should go back to spending hours formatting PowerPoint slides or manually compiling research that technology can now do in seconds.

What we're less sure about is whether we've thought through what disappears alongside those tasks.

We were reading PwC's latest AI Jobs Barometer recently. One statistic really stayed with us. After analysing more than a billion job adverts globally, they found that entry-level roles most exposed to AI are now far more likely to require skills such as judgement, leadership and strategic decision-making than they were only a few years ago.

That struck us as odd. We're increasingly expecting people to arrive with capabilities that previous generations developed by doing the very work we're now automating.

Perhaps that's fine. Perhaps organisations will redesign how people learn. But we're not convinced we've really worked out what that looks like yet.

Judgement has never come from reading a framework. It comes from seeing good decisions, bad decisions, political decisions and difficult decisions play out over time. It comes from making mistakes, defending recommendations, being challenged in meetings and slowly building an intuition for what matters and what doesn't.

That's much harder to accelerate. One of the reasons strategy has always fascinated us is that, despite all the frameworks, models and analysis available to us, there has never been a formula for good judgement. Two people can look at exactly the same data and come to completely different conclusions. Experience, context and intuition still matter enormously.

We suspect that becomes even more important in an AI-enabled world, not less.

If everyone has access to increasingly similar tools, then the advantage won't come from producing another report five minutes faster than your competitors. It'll come from asking better questions, challenging assumptions more effectively and knowing when not to follow the obvious recommendation.

We need to think much more deliberately about how organisations develop judgement once that work has gone. If we don't, we may discover that we've become incredibly efficient at producing answers, while quietly making it much harder for the next generation of leaders to learn how to think.