Back in 2007 when NICE Systems acquired Actimize I wondered how long it would be before someone figured out how to use the company’s fraud detection systems in contact centers. The answer is now, as NICE Systems announces its Contact Center Fraud Prevention system.
The original Actimize system used artificial intelligence techniques to track potential fraudulent financial transactions. It works by tracking financial transactions and consulting rules, including user, origin and transaction pattern information, to identify transactions that fall outside the rules. It can flag these suspicious transactions and create alerts so a transaction can be stopped and someone can investigate.
The new contact center system works on similar principles. It begins by using voice biometrics to create a voice profile of callers, combining physical and behavioral factors to create a unique and content-independent voice pattern for each individual. It then uses real-time voice recognition to match a caller to a list of suspects on a watch list. It uses a combination of speech analytics, emotion detection, speech patterns and information captured at the agent’s desktop (such as customer ID, past interactions and topic of call) to determine a risk score for the interaction. This is then combined with contextual data (such as calling number, IVR responses and origin of call) to assess the risk of the call. Calls at risk are processed by the rules engine, which can alert the agent or other individual of the potential risk and say what action to take. It also creates a fraud case that can be used to manage further investigation.
The same process can be used to make the caller authentication process more efficient and effective. Real-time voice recognition can match the caller’s voice profile to a list of stored profiles, and if a match is found (and the caller has not been identified as someone at risk) then the call can be allowed to progress. If no match is found then the system can use other forms of analytics on call information and previous transaction data to determine the next action for the agent. Trials indicate a success rate greater than 95 percent, so such a process can speed up authentication, make it more secure and improve the customer experience.
The interesting thing about fraud is no one ever knows how bad it is until it is detected. This new product from NICE Systems allows organizations to begin this investigation and where needed put in place new processes to reduce potential risks. Two potential downsides of the software are getting it wrong and annoying genuine callers, and possibly falling foul of regulations. However, used with these potential pitfalls in mind, the Contact Center Fraud Prevention system offers organizations the opportunity to close a door before issues occur. I recommend companies investigate whether their business justifies the use of such a system.
Richard J. Snow
VP & Research Director