The Skills You Stopped Buying: Where Judgement Actually Comes From

Entry-level work is disappearing from British organisations. The British Chambers of Commerce set out the scale of it last month, alongside a forecast that youth unemployment will reach seventeen per cent this year.

Automation is part of that and it is not all of it. Which makes the consequence rather easier to discuss than the cause.

1. It Is Not Only Automation

The temptation is to attribute the whole thing to AI. The evidence does not support that as neatly as the headlines suggest.

NIESR has estimated that changes to national insurance, the minimum wage and employment rights raised the real cost of hiring an entry-level worker by around seven per cent. Several large employers have named the economy rather than technology as the main factor behind a reduced intake.

The causes are mixed and will probably stay that way. The effect is not in dispute — fewer people are coming in at the bottom, and the tasks they used to do are being done some other way.

2. Nobody Decided to Stop Developing People

It is worth being clear that this was not a decision anybody took.

No board resolved to reduce its supply of future senior staff. A series of separate and sensible choices — automate this, outsource that, defer a graduate intake by a year — added up to one.

Each was defensible on its own. The aggregate never appeared on an agenda.

3. Judgement Was a By-Product

What is being lost is hard to name, which is a large part of why it goes unnoticed.

People acquire judgement by spending years on unglamorous, high-volume work. Reconciliations that will not balance, first-draft analysis, checking somebody else’s figures. Nobody set out to teach anything — the work had to be done, and doing enough of it produced somebody who could tell at a glance when a number was wrong.

The organisation was never paying to develop anyone. It was paying for the work, and the judgement arrived free with it.

4. The Saving Is Immediate. The Loss Is Not.

Which is what makes this so difficult to govern. The cost reduction lands in the current year, and it is real money.

The consequence turns up around year seven, when the organisation needs somebody who can look at an output and know it is wrong before being able to say why.

By then nobody connects the two. The saving was approved by a person who has since moved on, and the shortage presents itself as a recruitment problem.

5. Reviewing the Output Needs What You Stopped Producing

There is a circularity here that deserves more attention than it gets.

The obvious answer is to keep people for the review — let the machine draft and have a person check it. But checking is the harder skill rather than the easier one, and it is the one that came from having done the work.

An organisation that automates the doing and keeps only the checking has retained the part requiring experience while removing the route to acquiring any.

6. None of Which Argues Against Automating It

This is not a case for keeping people on work a machine does better. That would be expensive, dishonest, and unkind to the people concerned.

Nobody should spend three years reconciling ledgers as a character-building exercise if the reconciliation takes seconds. The work itself was never the point.

The point is that it was quietly doing a second job nobody had costed — and that second job still needs doing.

7. What Has to Be Deliberate Now

Which means designing something that used to happen by itself. Four questions cover most of it:

• which judgements the organisation cannot buy in, and must grow
• where somebody at two years’ experience now gets the volume to form them
• who is accountable for that pipeline, and against whose budget
• how long it honestly takes to replace one of those people from outside

The last turns this from a values conversation into an entry on a risk register, which is where it will get taken seriously.

8. It Has Started to Cost Something

The uncomfortable part is financial. Development used to be free, because it rode on work that had to happen regardless.

It does not any more. Producing the same judgement now needs deliberate spending — rotation, exposure, time alongside people who have it, work handed to somebody slower than the alternative.

That is a real cost against a benefit arriving years later, which is close to the least appealing shape a business case can take. It is also unavoidable.

Final Thought

Every organisation says its people are its greatest asset, usually in a document nobody reads twice.

The more useful question is narrower and answerable. Where exactly will the person who spots the expensive mistake in 2033 be learning to spot it — and is anybody paying for that this year?