I was fortunate and fascinated, nearly a decade ago, to meet Sir David Rowland, saviour of Lloyd’s. For readers who don’t know the name, or perhaps aren’t aware of the gravity of the precarious position of Lloyds in the early 1990s, suffice to say that the market was set to collapse before its 300th birthday, and Sir David led the team which brought it back from the precipice.

Sir David, who passed away in 2019, was a broker. He began his career in 1956 as a junior with Stewart Wrightson, where, he once told me, he spent the first two months of his tenure with a pot of glue. His job was to paste exclusion clauses printed on paper the thickness of onion skins onto the inside of policies. The job required care, he explained. “Too much paste, and the whole policy would be ruined, stuck together in a gluey wad. Too little, and the exclusion would come fluttering out.”

The paste was made from fish scales, lending Sir David an eau de poisson for his entire market debut, but he died with the Lloyd’s Gold Medal, a knighthood, and the unyielding respect of everyone whom he ever encountered.

In late September three decades after Rowland’s Reconstruction & Renewal programme saved Lloyd’s, Computing Magazine published news I found interesting, if not predictable. “Accenture is laying off more than 11,000 staff that CEO Julie Sweet claims cannot get with the AI programme,” journalist Graeme Burton wrote. With such news, it’s starting now to feel that the AI revolution is a lot more than just hype. Real people’s livelihoods are on the line.

But there’s another part to the Accenture story. Sweet positions the layoffs as happening in the context of Accenture as no company for old men. She insists that despite losing the 7% or so of her staff who don’t do AI, the firm’s headcount will increase before the end of 2026, with a focus on the AI-abled. Accenture will be training everyone, from school-leavers to their most senior consultants, to make the most of technology day to day.

Insurance leaders can learn from the example. A report by Economist Impact entitled ‘Underwriting the future: the role of artificial intelligence in insurance’ presents an AI-focused talent risk for our sector. “As technology automates tasks, employees’ roles will evolve and require skills training in areas where human input remains critical,” the article declared based on panellists’ points.

So far so good. But the risk? “There was a debate among experts on whether the next wave of agentic AI would replace certain roles or change their nature. In any case, this transition won’t happen overnight: there will be resistance to change – especially from actuaries, underwriters and lawyers – and upskilling will become challenging if technology renders junior positions redundant.”

Companies that avoid this pitfall will reap the benefits of AI. Others will go the other way. Training juniors will be difficult if AI makes them redundant, but not impossible. We need simply to find to rethink training while embracing the technologies that will allow our businesses to grow. And we need to consider the old pros, as well as the newbies.

Agentic AI, for example, holds out great promise to make the insurance value chain more efficient. It can free-up the valuable time of skilled professionals by saving them from mundane tasks, but they need to know how to use it efficiently. AI should and will play a big role in the elimination of the mundane tasks done by juniors (data rekeying, for example, is a function which should be banished from our market), but we will have to find some other way to show them the ropes.

We frequently decry the hollowness of our talent pipeline. If we leap at the short-term bottom-line enrichment AI proffers by replacing juniors without rethinking how we train our people, we risk handicapping our future. By eliminating the cut-and-paste jobs (whether literal or virtual), we risk sacrificing the learning that makes great mid-career professionals and future leaders.

We must re-think training at every level to avoid the nose-to-spite-face dilemma. When NEDs and C-Suites consider where AI can be deployed to make their organisations more efficient, they must also think about training. The next David Rowland won’t come up through the ranks with a glue stick in hand.

Guillaume Bonnissent is CEO of Quotech.