By submitting your email address, you agree to receive emails regarding relevant topic offers from TechTarget and its partners. You can withdraw your consent at any time. Contact TechTarget at 275 Grove Street, Newton, MA.
Richard Davies, the IFB's deputy chairman, told Computer Weekly in an exclusive interview that the results "exceeded all expectations".
The IFB has outsourced its data mining operations to Detica, a specialist IT company. Its NetReveal software applies social network analysis to huge amounts of data to identify, understand, and evaluate higher-level networks of potentially collusive individuals and organisations.
The IFB uploads claims data daily to Detica which then uses data items such as names, addresses, birth dates, claims history and tip-offs to identify relationships and their relative strength between claimants.
Detica ran a proof of concept pilot for the IFB in 2005. "We never expected it to be as powerful as it has turned out," said Davies.
The IFB has used Detica's services initially to uncover gangs that cause or fake car accidents to defraud insurers. Passing leads to police forces and the Serious Organised Crime Agency has led to 74 arrests and "major disruption" of several gangs, Richards said.
"I have been in the business 17 years, and I cannot remember a year when police made more than 10 arrests for insurance fraud," he said. However, none has been prosecuted yet because police and the Crown Prosecution Service are still preparing cases.
But the Detica leads have allowed police to seize goods worth £5.5m under the Proceeds of Crime Act. This could be used to repay insurers who have lost millions to fraudsters.
Davies reckons the arrests have saved insurers at least £8m. "We reckon insurers lose 5% of premium income to fraud," he said, "so Detica is helping us to reduce that loss."
Whether these lead to lower premiums remains to be seen. However, Davies believes it will be easy to sell an expansion of the service to his 46 members and to rejustify his £8.6m, five-year budget.
Comment on this article: firstname.lastname@example.org