Voice risk analysis has made the underwriting process more efficient. Richard Sheridan explains how this technology and human judgement can work together in other arenas.
The insurance industry has been transformed through the use of technology; it can now provide instant quotes and cover over the phone and internet. The claims process is also increasingly automated. Ease of use and efficiency has benefited customers and insurers but how does this impact on the human element of the process? In maximising the ability to improve service and efficiency, it is important not to lose sight of the value of skilled decision making.
The use of voice risk analysis (VRA) technology to support staff decision making was pioneered by UK insurance firms, saving millions of pounds in claims costs and improving the handling speed for genuine customers. It is now being piloted by the Government and 25 local authorities for use in benefits claims and its initial success is attracting worldwide interest.
For example, Birmingham City Council was able to reduce the time taken to process changes in benefit entitlement by 33%. Using the technology, it verified the information provided in around 65% of cases as ‘low risk’ and waived the requirement for paper verification of the information in around half of those cases. This saved the council 33% in service costs. Perhaps more significantly, a customer survey showed that 98% of customers were either satisfied or very satisfied with the process.
Though most current public interest is in VRA’s deployment in the benefits system, it is important to remember that the UK insurance industry led the way and now further developments are planned in claims and underwriting, which will help to identify and manage risk before there is a chance for a claim to arise.
The software measures the minute fluctuations that occur in people’s voice frequencies as a result of their emotional reaction to the situation or what they are saying. The software is able to interpret these tiny, inaudible variations and provide an insight into how they respond to statements. This ranges from ‘low risk’ at one end of the spectrum, through indications such as ‘uncertainty’, ‘excitement’ and ‘tension’. All of these are natural reactions to what we are saying and do not give cause for concern.
Warning signs
Where there is cause for concern about a particular phrase then the technology will alert the user with a message that identifies ‘high risk’. This is therefore an immediate and objective signal to users that there may be a problem. This is where the importance of human judgement is vital.
Though the technology will alert the user to ‘high risk’, it is not recognising what is actually said, just the reaction to it. It is therefore vital that a person makes a decision about whether the risk is relevant to the scenario involved. People’s motivations for being untruthful are numerous and, as part of a risk assessment process, the trained member of staff must make a decision on what is a relevant risk.
The key to the VRA process is in combining the capability of the technology with intelligent process design and staff training so that decisions are taken under the control of staff to enable each case to be assessed on its own merits.
The use of the process in insurance claims is likely to expand to include third-party personal injury claims. For a long time, the length of time and costs involved in settling these alongside their associated fraud risk have been a source of concern for insurers. The impact of Ministry of Justice reforms is likely to be a requirement for speedier decision making and VRA should assist in this.
Underwriting is also an area where VRA is already being deployed. Detecting risk at the point of inception will benefit insurers in two key ways. Firstly, the process can be used to provide an enhanced level of certainty around the information provided by genuine customers – enabling underwriters to confidently target these people with competitive rating. Secondly, the capability to identify concerns will allow underwriters to avoid writing undesirable risks in the first place, which will reduce the exposure to claims and the cost of handling and investigating these.
One final area where VRA will develop is in its benefits for staff. VRA does not replace staff but empowers those traditionally seen in a purely processing role. The impact of staff retention has been positive with a 75% improvement at one insurer. Secondly, the process is now being developed to assist insurers and other organisations in the recruiting of the right staff. The ability to assess risk in the answers to routine pre-employment questions will help to weed
out prospective employees where there is concern about the accuracy of the information provided.
Insurance is the risk business and for an industry often accused of lagging behind in technological matters, its continuing development of the use of VRA is a great example of innovation now being taken up by the Government and industry sectors, both here and abroad.
Richard Sheridan, client services director, teceris validation services, Capita
