Why AI must not become an excuse to stop hiring junior staff

They say history doesn’t repeat itself, it rhymes, and that certainly is the case within the insurance profession. 

Currently, it is abuzz with discussions about how to safeguard its future talent pool amid a slowdown in graduate hiring. It is a trend accelerated by artificial intelligence taking over entry-level roles. 

Given the substantial number of experienced professionals nearing retirement, neglecting entry-level recruitment risks worsening a long-standing skill shortage.

However, the reality is that the talent pipeline crisis existed long before the advent of generative AI.

In fact, CII research puts the proportion of young people who see insurance as a viable career at a mere 4%. Around one in four insurance employees are over 50 and more than 70% of insurers report shortages in data and digital skills.

Talent attraction and retention has risen from seventh place to become UK insurers’ top business challenge this year. AI is not the cause of that, but it could make it worse.

Hiring risk

Historically, firms justified avoiding entry-level hiring by pointing to a lack of time for training, seeking candidates ready to deliver immediate results under heavy workloads. 

Productivity enhancements from automation challenge this reasoning. 

The risk now is that firms treat the capacity released by AI as a reason to hire a tool instead of a person, when it is precisely the headroom needed to sit with a junior colleague. 

Given the substantial number of experienced professionals nearing retirement, neglecting entry-level recruitment risks worsening a long-standing skill shortage.

Engagement

AI also has the potential to streamline recruitment and boost attraction. 

By taking over administrative burdens such as keying, rekeying and chasing tasks, AI shifts the focus towards client engagement, negotiation and handling complex risk work much earlier in a career. 

If every firm competes for the same shrinking pool of experienced people, the cost of that experience rises until hiring stops being economical.

The insurance profession has long struggled to convince school leavers that it is a better choice than other sectors. 

That is not something AI solves on its own, and the same reduction in low-skilled administration is happening across the wider economy. 

What it does mean is that the work we can offer a new entrant is more interesting and engaging than it once was, and that is a case worth making far more loudly than we do.

Judgement at risk

Development, though, has to become deliberate, because judgement used to be a by-product of volume. 

A decade ago, you learned wordings by navigating them: how they are structured, which clauses to expect for a given trade, what is conspicuously absent. Put the wording through a tool now and you get a summary. 

An experienced broker reads that summary and recognises broadly what they expected to see. A junior colleague has no such benchmark. 

So, an omission or hallucination in the output carries an increased risk of going unchallenged. AI accelerates development where the trainee can verify the output and quietly erodes it where verification requires judgement they do not yet have.

Part of the answer could be to consider more carefully how the tools are used, by encouraging trainees to form their own view before running the tool and then comparing the two. 

Used that way, the tool becomes a marking scheme and they build their own understanding rather than borrowing it. Another aspect is ensuring there is always a human in the loop.

None of this works without proper oversight, monitoring and coaching from senior colleagues who do have that knowledge and experience, and who are close enough to the work to notice when something needs correcting.

Apprenticeships should be considered as a route that gives that kind of development a structure. From 1 August 2026, training for eligible under-25s at non-levy-paying employers is fully government funded. 

That means cost is no longer a credible objection for smaller firms, and programmes can be built around progression and retention rather than short-term cover. Alongside that, any firm using these tools should have a governance framework setting out where AI may be relied upon and where a human must verify.

Shared obligation

The obligation sits with everyone. Responsibility lies across the entire sector, spanning brokers, insurers, and managing general agents alike. 

While many larger firms may already have structured early-careers routes, there is an opportunity for the long tail of smaller and mid-sized firms to open more doors to younger talent pools, where the capacity released by these tools makes doing so realistic perhaps for the first time.

One perspective is that the market will eventually force the issue. If every firm competes for the same shrinking pool of experienced people, the cost of that experience rises until hiring stops being economical. 

At this point, firms will have no choice but to recruit less experienced staff and train them properly. The question is whether we arrive there by design or by hitting the wall. The firms that wait will be buying experience at the top of the market. 

The ones that train now will have grown their own. A profession that stops recruiting people who cannot yet do the job has stopped replacing itself.