The AI talent trap – and how to avoid it

Is that trade-off genuine? Or are we looking at a confluence of challenging economic factors during which businesses' use of AI happens to be gathering momentum? 

The Office of National Statistics reported in May 2026 that more than nine in 10 businesses using AI said there was no change to their overall workforce size. Our own data, gathered as part of our C-suite barometer 2026 research, confirmed that AI has had an impact on organisational structure, but not yet in the form of large-scale job losses. On the contrary, AI has had a net positive impact on jobs – while 27% of leaders say AI has already replaced jobs in their organisations, another 71% say it’s actually created new roles.   

Of course, not all roles are the same, and it also appears to be the case that entry-level or admin-type jobs are more likely to have been affected; even then, employers report using AI to handle certain tasks, rather than replacing roles outright. This follows the principle of keeping humans in the loop and using AI to augment rather than replace teams. So far, the story seems to be one of redeployment, augmenting and retraining rather than large-scale removal. 

The narrative oversimplifies the reality

If the machines aren’t quite taking over yet, there might at least be an argument that graduate or entry-level roles are most at risk as companies learn more about how AI can be used. 

Asam Malik, technology partner and UK executive board member for Forvis Mazars, says that while AI has much to offer, companies need to carefully manage their organisational structures to avoid talent bottlenecks in future. 

"AI has been a focus on board agendas for a while now, but the bigger concern for companies at the moment is the economic pressure that’s being felt not just here in the UK, but across the world."

Diamonds may not be forever 

The traditional ‘pyramid’ structure of organisations – with larger numbers of graduate and more junior levels at the base, and decreasing numbers of more senior people towards the top – is under threat from any emerging ‘AI as the foundation’ orthodoxy. There may be short-term cost savings in a new diamond-shaped structure, whereby a stronger mid-level of experienced people represents the greatest proportion of the workforce, and the traditional base is handled by a smaller number of people working with AI.  

But Asam points out a risk in this approach. With fewer graduates, the future of the organisation changes. Most organisational models rely on something we might call managed attrition. You hire larger numbers at the entry level, expect a significant number to leave, and you focus on retaining the right people to form the next layer of leadership as they progress. 

But if the intake is smaller to begin with, attrition doesn’t self‑correct. You end up with a thin middle, fewer future leaders, and far less resilience than you expected. 

This, then, is a talent issue, rather than an AI issue. 

Retention as an operating model  

Which brings us to the convergence of AI and talent. If it’s true that, to some extent, junior hiring will at least change shape as AI adoption becomes more established, then leaders will have to adapt their talent strategies to meet the impact of this. 

In other words, if you’re hiring fewer people, you need more of them to stay. Development, progression, and experience have always been part of recruitment and retention strategies, but the ‘diamond’ model promotes them to existential necessities.  

That requires different investment decisions, different leadership behaviours, and a different level of honesty about what the organisation actually needs from its people. 

“AI is already changing how we work and we’re still in the first chapter of what that’s ultimately going to look like,” says Asam. “What we mustn’t do is get into the habit of blaming technology for the hard decisions that are sometimes forced on businesses by economic circumstances. That can make it harder to implement measures which will benefit the business into the future. 

“What is clear, though, is that organisations are going to have to adapt more than just the technology stack  Only an organisation-wide strategy for adopting AI will enable  organisations to realise the benefits, in a way that doesn’t inadvertently hamper productivity further down the line.”

Get in touch