Zurich fraud record: £109.3m is not just a detection win

£109.3 million is the number Zurich UK wants the market to notice, and I think it is the wrong number to stop at.

Insurance Business reports that Zurich detected that sum across 4,359 fraudulent claims in the twelve months to 30 September 2026, its highest year by value and volume, about one bogus claim every two hours. Casualty took £61.6m, motor £28.1m and property £19.6m. Connected AI and analytics, built with Quantexa, found 860 of those claims worth £17.9m, roughly one in five of the cases Zurich caught.

My reading is that a rising detection total is evidence of both better tools and a fraud book that was never as small as the industry priced it.

The industry backdrop makes that harder to ignore. The ABI put detected fraudulent claims at £1.34bn in 2025, up 14 per cent even as case numbers fell. Fraudsters are aiming at bigger payouts. Zurich’s Scott Clayton, who also chairs the Insurance Fraud Bureau, told Insurance Business that networks are adapting with staged accidents, ghost broking, fabricated claims and AI-generated evidence. Insurance Age reports the same period and the same £109.3m total.

What the figure hides is the fraud that still lands as a paid claim. Detection is not the same as prevalence. Zurich’s own story is that linking Claims data to Companies House, Equifax, industry intelligence and the Insurance Fraud Register exposed aliases, shared accident locations and rogue solicitors. That is useful. It is also a reminder that siloed claim files were the industry’s default for years.

The same analytics, Zurich says, help pay honest claims faster when the fraud score is low. That is the part claims teams should press for. A tool that only slows the bad cases is half the job.

If your fraud detection total is rising, are you celebrating the tool, or revisiting the rate that assumed a quieter book?