Beating the insurance scammers

You get home from work and find the house in even more of a mess than usual. Except the kitchen window wasn't smashed when you left this morning. It doesn't take you long to realise that all your CDs, DVDs and your iPod have been stolen. Still, as you inspect the damage you figure the insurance will pay.

And then you think - insurance is so expensive, why not add a few extras to get a bit of that money back. Maybe tell the insurer the camcorder got pinched too? They'll never find out...

It's a tempting thought - but increasingly would-be fraudsters are finding out the hard way that insurers are using technology to crack down on dodgy claims.

According to the Association of British Insurers around £3.5m in insurance fraud is spotted every week. But it's hard to put a figure on the amount of fraud that is slipping through the net.

What is clear is that people don't see ripping off insurers as a particularly shocking crime.

According to research conduced this summer by consultant Accenture, 16 per cent of respondent said they knew of someone who had submitted an exaggerated claim to their insurance company.

A quarter of these dodgy claims were to do with motor insurance, and more than half were property related. More than half (53 per cent) of those surveyed said it was acceptable to inflate a claim by £100 to cover some of the cost of the insurance. And most people wouldn't report a neighbour for making a fraudulent claim.

With this level of fraud, it's not surprising insurers are beefing up their defences. One technology being used to fight back is data matching, which is especially effective against people who make fraudulent repeat claims - click here to read a case study featuring insurer MMA. Many of these claims are made by organised gangs.

But Steve Lathrope, a partner in Accenture's general insurance business said gangs are easier to spot than exaggeration by the average claimant.

He said: "What's harder to address is the inflation of individual claims."

Lathrope said: "The reason it's hard to detect is there is no point data-matching against it because a customer might only do it once."

The volume of claims being made is another problem here: "Typically claims departments are high volume operations taking lots of calls. That's not a good environment for stopping fraud because they are more focused on providing good customer service and moving on."

In response to this some insurers are applying more business rules when they get a claim, he said.

For example, a claim for an expensive camera stolen from a car late at night, just after the policy had been bought, might be flagged by the system as a case that needed more examination. Sometimes just asking more detailed questions is enough to make people withdraw their claim.

Another way technology can help is to use voice stress analysis, which is being utilised by a number of UK companies, including eSure, Halifax and Highway Insurance.

DigiLog's fraud prevention process - used by those insurers - uses a combination of Voice Risk Analysis technology with careful questioning to spot and deter fraudsters.

DigiLog's managing director Kerry Furber said its offering combines technology and tailored questions - and its findings suggest that as much as 15 per cent of claims include some sort of fraud: "What we find consistently with our customers is that on average 30 per cent of claims indicate there is risk within them and we defeat about 50 per cent of them. Genuine customers will engage with you but fraudsters want to give away as little information as possible."

There's even a final tweak - if your fraudulent claim makes it through all these hoops, you might find you get paid with an electronic voucher, not cash.

Argos Business Solutions supplies many insurers with cards which they use to pay claimants. Genuine claimants often simply want goods replaced quickly, and the cards can be programmed with an item's Argos catalogue number if a specific item has been stipulated. Fraudsters looking for a cash payout will be disappointed.