Voice Biometrics Improve Transaction Monitoring Fraud Detection
Transaction monitoring fraud detection systems are now commonplace at financial institutions. But what can be done to improve and prepare them for the future? One answer is to extend them to integrate with new voice biometric analysis systems.
As a result of fraudsters evolving and altering their behavior, legacy transaction monitoring systems are known to miss some fraudulent transactions because they are consistent with historical, legitimate transactions. What are some of the latest technologies being applied to improve and update this situation?
By integrating voice biometric data into the process, we can help improve fraud detection rates. This is done by adding a comparison of the voice used to initiate the fraud scheme or the fraud transactions with the legacy transaction analysis. Additionally, voice biometric analysis can assist with the post-fraud transaction verification processes, so calls to fraudsters who confirm their fraudulent activity as "legitimate" are detected.
Fraudsters in wire transfer fraud and card fraud schemes are known to have researched the legitimate customer's prior transaction activity in order to make their fraud transactions look as normal and customary as possible. For example, fraudsters are known to falsely generate wire transfer requests that are only slightly different than previous legitimate wire transfer requests. Similarly, fraudsters are known to mirror prior card activity with "normal" merchant category codes, transaction amounts, and transaction geography on fraudulent card transactions.
It is no longer enough to just evaluate such elements as the amount of a wire transfer, where the wire is being sent, when the wire is authorized, and to whom it is being sent. For phone-initiated wire transfers, you should also include analysis of the voice requesting the wire. Voice biometric analysis and voice prints are already successfully being used to identify both known fraudsters and legitimate customers.
As with any predictive detection system (transactional or voice), these are probability-based predictions. By combining moderate- or low-risk transaction scores with moderate or higher scoring voice print analysis, the combined assessment can indicate high risk. The resulting detection derived from combining the two risk assessments is far greater than it would have been using the individual assessments.
The types of activity that provide the opportunity for combined voice-transaction analysis are not unknown or even esoteric. These familiar activities include a long list of normal events: voice requests for password resets, change of address, card reissues, check orders, wire transfer requests, adding additional signers, replacement statement requests, registering for online banking, and even fraud confirmation calls made to the “customer” in account of take-over fraud cases. All of these are perfect applications for combined transaction and voice analysis. Frankly, any conversation between the consumer and a financial institution's call center, back office, or branch that can further a fraud scheme can be an opportunity to integrate transaction scoring with voice scoring.
Whether it is distributed denial-of-service attacks staged to drive volume to backup call centers in order to hide fraudulent call center requests, bogus telephone wire requests, or the expected increases in social engineering and account take-over, there should be little argument that it is time to enhance existing transaction detection systems by integrating voice biometric data analysis.
And the signs indicate that financial institutions are already on board. In an Aite Group report, when financial institutions were asked whether they plan to invest in voice technology as one method of protecting their call centers, 25 percent of them indicated that they either have a pilot underway or are working on a production rollout. Another 40 percent indicated that they have voice technology on their one- to two-year call center roadmap.
Clearly, there is little argument that it is time to enhance existing transaction detection systems by integrating voice biometric data analysis.
Wesley Wilhelm (Wes) has more than 30 years of experience in banking and consulting to the financial services industry, with extensive knowledge of fraud management, payments, and retail banking technology and operations. He has held numerous management positions in risk and ... View Full Bio
Tweet This