Voice stress analysis may help in fraud detection, but it may also damage the industry's public image rather than contribute to reducing the annual £2.2bn fraud bill.
Two industry heavyweights argue the pros and cons.
Kerry Furber, Highway head of fraud risk management
StuartCliffe, National Association of Bank and Insurance Customers chief executive
1. Do you think voice stress analysis (VSA) can make a serious dent in the
£2.2bn annual fraud bill?
Not on its own. In our forthcoming trials, VSA will play a supporting role to a much broader cast of contemporary techniques - all providing significant assistance in the fight against fraudulent claims. While VSA can greatly enhance the ability to spot truth or deception,it is only when deployed alongside these other powerful but, in the main, non-technological solutions that it helps to tip the scales of justice hugely against the fraudster.
The combination of techniques are designed to enable swift identification of genuine claims which can then be moved on for fast track settlement - avoiding the often frustrating mechanism of less scientific, traditional insurer's practices.
Any perceived reduction in claims payments will be mainly
artificial. Customers who have mobile, cordless and hands-free phones, suffer from speech impediments, who don't like to make telephone calls, happen to have a cold, or are under stress when they make the call, can presumably look forward to having an even harder time getting claims paid by underwriters. Delays while these cases go through the complaints process and to the ombudsman or litigation should nicely improve claims cash flow in the short term. Longer term, more cases will go to the ombudsman and to litigation (where VSA does have some slight drawbacks) and a code of practice will be hammered out as to exactly where and when - if ever - VSA is appropriate. Meantime, the industry will be spending even more money
a) giving customers a reason to dislike providers and b) rushing to use the wrong tool at the wrong time.
2. Do you think VSA could be perceived as an infringement of civil liberties?
Currently, it is the civil liberties of genuine policyholders that are being infringed by the increasing numbers of fraudsters whose activities dictate that everyone else must foot their bill. It needs to be understood here that with the Association of British Insurers (ABI) now estimating at least 10% of motor claims as fraudulent, we would be failing in our duty to our stakeholders, including genuine policyholders, if we did not consider and test new and emerging methods and technologies that can assist in reducing this exposure.
Not only perceived as an infringement, VSA so far appears to have no standing in a court of law. Its major practical application so far has been where individuals are actively suspected of some act, and can be intimidated in a controlled, usually face to face situation, into making mistakes, which may lead to a confession or at least other evidence which then supports a conviction. The industry presumably intends to use publicity about VSA monitoring of calls to persuade customers not to lie or exaggerate claims.
3. Do you think VSA has been subjected to enough testing and scientific research?
I have been making enquiries about VSA technology for almost four years. The specific technology we are considering has been subjected to a number of academic studies and all of the ones I have read show it to be highly effective in detecting deception (80% to 90% accurate). It is worth pointing out that VSA has evolved significantly since most of those studies were undertaken.
If we go on to consider that, currently, the average insurer is almost certainly less than 1% effective in spotting the frauds using traditional methods, VSA clearly represents a giant leap forward. Further, I have seen it working highly effectively in a variety of test and live situations and conclude that only a latter day Luddite would ignore its potential.
My research seems to indicate that VSA divides experts quite significantly. Those `for' agree that it is far more effective, portable and convenient than any other physical test, although there is little information available on telephone use. Those `against' and there are some influential forensic names in that category - regard the application as being in the same category as tea leaves, dowsing and gut feeling.
A more appropriate question might be: Does the customer believe that VSA has been subjected to enough testing? Any customers wrongly accused of fraud through VSA will fight with a righteous indignation that I suspect may surprise many insurers.
There is a difference in being told `you don't have a receipt' to `we know you're lying'. The first high-profile embarrassments for insurers will no doubt have wide publicity.
4. Do you think claimants should be made aware beforehand that they are being subject to VSA?
For much of the past, claimants have been largely unaware of what may have formed part of any process instigated to validate their claim. Additionally, we have only previously been able to take a less than educated guess as to whether a claim was suspicious or not, resulting in the possibility of both genuine and cheating claimants being subjected to scrutiny. We will be informing customers who ring in to our call centre that all telephone calls relating to applications and claims may be tape-recorded and the recordings used for fraud prevention and detection. We will be reinforcing this message in all relevant documentation and throughout the key areas of the claims process.
The research material I have been able to find places great importance on being able to ask test questions where the operator knows the answers are false, to compare the reaction to those carefully worded questions to which a reply is unknown. It seems likely that callers will have to be informed that VSA is being used in order to get a useful reaction but detecting damaging admissions seems more likely to depend on the importance of the situation to the individual. This would seem to indicate more successes against private individuals with a house or a car to lose, than organised fraud where a practised liar with some technical help might pass the test with flying colours.
5. Do you think insurers should be allowed to refuse a claim on the back of VSA results?
On the basis of VSA alone? No, not unless there was a degree of certainty in a VSA analysis that showed it to be accurate `beyond reasonable doubt', as per the burden of proof in criminal cases - thus making the test result `real evidence' in its own right. That may
happen in the future, but I think we are some way off this at the moment. All claims declinations should be made on a case by case basis and according to the evidence available.
Absolutely not, although claims are turned down without full (or any) reasons being given, and customers might not know that a claim has been rejected because of a `failed' VSA test. The uncertainties over these tests, including the details of their conduct and degree of reliability, mean that any decision to turn down a claim based solely on VSA will be purely arbitrary. Unless convincing additional proof is available, VSA should be an indicator only.
6. What other methods of combating claims fraud might be more socially acceptable by consumers?
Apart from a banner headline in an insurance press article, with no supporting research - or even views from members of the public contained within the body of its report, I have yet to see any evidence at all that suggests VSA as being socially unacceptable to consumers. However, if we accept that at least 10% of claimants are fraudsters,it follows that this section of consumers would seemingly not welcome it. Despite the lack of research, I can state that since Highway publicly announced the intended use of VSA, we have attracted a good deal of publicity worldwide and are pleased to say that only one person has contacted us or the ABI about its use. But far from decrying VSA as socially unacceptable, the person believed that she was the unwitting victim of a staged accident and asked whether we could interview the other driver using the `lie-detector' she had read about in the paper.
Insurance needs to become more socially acceptable to its customers. The industry has a Scrooge-like image with serious question marks over the value of its products, the honesty of its management and sales processes,and the efficiency and fairness of claims settlements. Insurers need to clean up their act, pay claims fairly and efficiently and provide a genuine customer service. Improved service for customers means increased communications, more background information and greater knowledge.
7. Could VSA lead eventually to a reduction in premiums? In short, it is the right combination of methodologies and technology that will haul insurance claims validation out of the Dark Ages and into the 21st century. From what I have seen, VSA very definitely has a role to play in this. For us it is in helping to quickly identify the genuine, for others it is in spotting deception. Either way, as I keep emphasising, VSA results cannot yet be used as the sole reason for declining a claim that could contribute to lower costs and subsequently drive down those premiums. So, we need
to stop thinking of VSA as a stand-alone decision making tool. It is a support tool that can help speed up what would otherwise be a pretty cumbersome and costly process. In this way I see it as having a significant role to play in the wider process that will help bring premiums down.
No.VSA will not be effective against major fraud, and will generate far more work with disputed claims. Some interesting legal costs may be incurred along the way.Any credible use of VSA would have to include substantial training programmes for operators and interpreters of results. It seems more likely to increase costs than make significant savings.
