This article is by Jimmy Williams, CEO and Co-Founder at Urban Jungle
Most honest customers have no idea they’re paying for other people’s insurance fraud; after all, it doesn’t show up as an added extra when they purchase a policy. Instead, it silently bumps up the price of every premium. But it does more than just add to customers’ costs; it directly influences who gets cover in the first place, and who doesn’t.
A tax that lands hardest on the people who can least afford it
Fraud is genuinely difficult to spot, and it isn’t spread evenly. It tends to cluster in lower-income areas, often linked to organised groups rather than one-off opportunists. Faced with that, a lot of insurance providers reach for the simplest lever to pull: they price up entire postcodes where fraud has occurred frequently in the past, or they refuse to quote them at all.
It’s an understandable shortcut, but it’s not a fair one. It punishes everyone on a street for the behaviour of a handful of people, and it does so regardless of whether any individual customer has ever made a dishonest claim. The result is that the people who are already the most financially exposed, the ones with the least savings to fall back on if something goes wrong, are often the ones locked out of cover altogether; or they are asked to pay higher premiums than they otherwise would.
Essentially, it’s a hidden tax. It isn’t paid by insurance providers, but by honest customers: in higher premiums if they’re lucky enough to get a quote, and by being rejected for cover altogether if they’re not.
Where AI actually earns its place
Fraud detection is where AI has made the biggest difference for us. Rather than solely relying on a customer’s postcode as a proxy for risk, we look at how people actually behave when they interact with us and manage their policies. We think it’s fairer to judge people on what they actually do, rather than on assumptions about where they live.
A genuine customer buys insurance in a fairly ordinary way; but someone with fraudulent intent often behaves differently (for example, buying a policy incredibly quickly), and we often see those unusual behavioural loops repeat themselves. We are using AI to tackle this by searching through unstructured behavioural signals and analysing patterns across the entire customer and claims journey. It’s rarely a single red flag that gives someone away, but a pattern or combination of small signals that, taken together, reliably indicate fraudulent intent.
We set a very high bar for acting on that, because making a mistake means unfairly penalising someone who did nothing wrong, and that’s not a trade-off we’re willing to make lightly. Done properly, this kind of approach can support far higher eligibility rates than postcode-based pricing typically allows, without taking on extra risk. It works because we’re better at spotting the difference between a genuine customer in a difficult postcode and someone trying to game the system. For me, this is where AI proves its real value: when it has a genuine, measurable impact on customers.
Why this is a fairness issue, not just a fraud one
It would be easy to frame all of this purely as catching fraudsters: insurance providers catching criminals, tightening up their numbers, protecting their loss ratios. But that misses the point of why it matters to ordinary customers.
Every fraudulent claim that gets paid out has to be funded from somewhere, and that somewhere is everyone else’s premiums. Over 440,000 cases of insurance fraud were recorded in the National Fraud Database in 2025, the highest number ever and a 6% increase from 2024.
For every postcode written off because fraud is too hard to police, a group of honest people pays the price for a problem they didn’t cause. Tackle fraud properly, using the right tools to catch the right people, and you can do two things at once: bring genuinely better value for money to honest customers, and open the door to people the market has quietly written off for years.
That’s a fairness question as much as it is a technology one. The real test is whether it leads to lower prices and wider access to cover. Financial inclusion and fraud prevention are usually talked about as separate topics. In practice, they’re two sides of the same problem, and solving one properly is often the fastest way to make progress on the other.
What the industry should take from this
Insurance fraud will never disappear entirely, and no single company will solve it alone. But treating it seriously, and treating it as a customer fairness issue rather than just a commercial tactic or a way to manage risk, is something the whole industry could do more of. The behaviour of others shouldn’t decide whether a person or a home can get covered, and for insurance providers, it is the fairest thing they can do for customers.
Be the first to comment