AI is insurance’s defining peril and its biggest opportunity

Andrew Johnston, global head of InsurTech at Gallagher Re, was the latest guest on the Voice of Insurance podcast.

Few people have watched the collision of insurance and technology as closely as Andrew Johnston. As global head of InsurTech at Gallagher Re, he has chronicled the sector quarter by quarter since its first report landed in 2017, a body of work Voice of Insurance host Mark Geoghegan calls ‘the standard literature’ on the subject.

So when Johnston says the conversation has fundamentally shifted, it is worth paying attention. The latest Gallagher Re quarterly report runs to more than 120 pages and digs into AI not just as an investment theme but as an emerging casualty peril in its own right. This episode unpacks what a decade of hype cycles tells us about where AI is heading and why Johnston, by his own description a very bullish, optimistic person, thinks the industry has more to gain than to fear.

A maturing market, not a frothy one

After a decade of dizzying peaks and bruising troughs, InsurTech funding has settled into something steadier. Johnston maps it neatly onto the Gartner hype cycle: past the peak of inflated expectations, through the trough, and now climbing the ‘plateau of comfort.’

What has changed most is the quality of the conversation. “The companies that are raising capital are generally coming to the table with very sound business propositions,” Johnston said. “Insurers and reinsurers are investing at unprecedentedly high rates at the moment, but they’re generally into companies that are supporting back office efficiencies, rules-based automation. It’s not speculative bets on the poster child of the moment.”

The grandiose pitches of a decade ago – claims that the industry was fundamentally broken, or that a tech-enabled balance sheet could capture a slice of a $7trn market and be worth $20bn – have largely disappeared. In their place: humility, use cases, and founders who bother to consult insurance people before they pitch.

Dot-com, meet dot-AI

Today, Johnston notes, around 95% of InsurTech investment is flowing into AI and founders feel they must signal it. He draws a sharp historical parallel.

“In 1999, companies were referring to themselves as sort of dot-com companies. We would laugh somebody out of the room right now if they said that they were a dot-com,” he said. “There is an evolution of terminology that goes alongside all of this.”

But he is clear that AI is more than a label. Unlike the parametric, embedded and blockchain waves that came before, AI sits on top of existing technology making it horizontal, pervasive and, he believes, far more durable. In time, he expects the AI tag to fade simply because it becomes assumed, synonymous, the way machine learning quietly disappeared into the background.

AI as an emerging casualty peril

The most striking section of the discussion concerns AI as a liability risk. Johnston expects a trajectory closely mirroring cyber’s emergence around 2017–18: existing D&O and E&O policies will sharpen their wordings, but a standalone AI liability class will ultimately stand on its own, with early movers such as Testudo and Armilla already in the market.

A subtle point cuts through. The biggest driver of loss today, Johnston argues, is not technology failing it is mismanaged expectations. “One of the biggest drivers of loss is not actually something specifically going wrong. It’s that people’s expectations have been mismanaged about what the outcome should be,” he said.

That ambiguity – where does everyday business end and delegated AI begin? – will largely be resolved in the courtroom. Litigation, both Johnston and Geoghegan agreed, is inevitable as experience races ahead of regulation. Encouragingly, the industry is leaning in rather than hiding, having learned hard lessons from silent cyber. As Johnston said, “There’s actually no value in shying away from this.”

There is even a neat irony at the heart of it: the same tools creating the exposure may help solve it. AI is already being trained to spot AI-generated fraud, from faked crash photos to hammer-smashed hail damage on roofs. “I think AI is going to come to its own rescue,” Johnston said.

The human stays in the loop

For all the disruption, Johnston is firm that humans are never going away. The winning model, he argues, is the enhanced underwriter and enhanced broker technology handling predictable, menial, rote tasks so people can focus on relationships, strategy and growth.

He also sees a genuine social dividend: lower costs and technology’s transcendental reach could finally help close the protection gap, extending cover to demographics long blocked by physical and economic barriers. “I just see it as a win for everybody.

Why this matters now

The thread running through Johnston’s analysis is that AI is not another poster child to be ridden and discarded. Because it sits on top of existing technology rather than alongside it, the leaders who treat it as durable infrastructure, not a passing cycle to be survived, will be the ones best placed to benefit. That means getting ahead of the liability question rather than waiting for the courts to define it: sharpening D&O and E&O wordings, watching the standalone AI class take shape, and recognizing that many uninsured losses today stem not from technology failing but from expectations that were never clearly set in the first place.

It also means holding founders and internal teams to a higher standard than the slogans of a decade ago. The credible proposition now embeds AI in a defined business outcome and keeps the human expert firmly in the loop. Handled that way, Johnston argues, AI is less a threat to the underwriter and broker than the very thing that frees them to do more valuable work, brings costs down, and finally starts to close the protection gap. For an industry that has spent ten years cycling through hype, that is a notably grounded and optimistic place to land.

Listen to the full episode

This conversation ranges far wider than this summary can capture from Black Swans and turkeys to the cyber arms race and the metaphysics of using AI to test AI. Listen to the full episode of the Voice of Insurance podcast.