Insurance companies are increasingly using technology to monitor voice patterns and provide clues to possible fraudulent claims. A pilot project involving seven councils has now concluded that the technology is also effective in picking out callers who make potentially false benefit claims.
The voice risk analysis technology spots changes in a caller’s voice enabling trained operators to decide whether a call is high or low risk and what further action to take. Initial results from the pilots have shown that the risk-ratings made using VRA have been successful. Of the cases rated as high risk, an estimated thirty per cent identified a change that would not otherwise have been recognized.
At the beginning of each call, the characteristics of a customer’s voice frequency are sampled in order to establish a benchmark, taking into account any pre-existing emotions. The benchmark is used during a conversation as a guide for analysing changes in frequency caused by changes in emotions.
Calls are handled by operators trained in questioning techniques and various forms of behavioural analysis which allows them to assess the level of risk in the conversation. In addition, VRA technology is used to support the operator’s analysis. The technology analyses changes in voice frequency and performs thousands of mathematical calculations, resulting in the identification of different categories of emotional content. The patterns associated with these categories allow the technology to identify genuine callers.
The intention behind the technology is to improve customer service by lessening the intrusion and cost of the review process, which currently is postal and visit driven. Claimants assessed as low risk are asked to forward verification of any changes, but they are not subject to any more intense review processes. This approach will allow for those identified as high risk to be subject to further action, such as a home visit or fraud investigation.
Funding has been provided for a further 15 local authority pilot projects to develop the technology.