Improving Fraud Protection With Voice Biometrics

Mark Lazar, Verint Enterprise Intelligence Solutions
Mark Lazar, Verint Enterprise Intelligence Solutions

Professional fraudsters are known for targeting contact centers in the financial industry. To address these systematic attacks, some organizations are responding with a new generation of voice biometrics technology that can silently detect known fraudster voices during incoming calls. The goal of these new systems is to increase fraud protection without disturbing the customer experience.

Fraudsters and the Contact Center
Traditional security questions don’t always prevent these fraudsters, who are skilled in social engineering and armed with stolen identity data. As a result, it’s important to consider guarding all channels and points of communication with your organization. However, industry analysts identify the contact center as starting place and also a security weak link exploited by fraudsters in the face of tightened online security.

Behind the professional fraudster’s success is the weakness of the security question paradigm. Today’s fraudsters buy stolen identities and leverage social networks to answer many security and out-of-wallet questions. Fraudsters are also skilled at socially engineering agents who have dual roles—customer service (making callers happy) and security (being suspicious of callers). They learn enough to pressure agents into security breaches by repeatedly and systematically calling. In fact, one study found that a significant percentage of total fraud calls are “repeat attacks” carried out by a small set of professional fraudsters.

Voice Biometrics for Fraud Detection
In response, forward-thinking contact centers at banks and card issuers are turning to voice biometric solutions to further their fraud detection efforts. The technology works by profiling and recognizing a human voice’s unique vocal characteristics, or “voiceprint.” To perform fraud detection at contact centers, the voice biometric system stores the voiceprints of known fraudsters in a database and screens incoming calls against the database to detect fraudster voices. Upon detection, it proactively notifies the live agent and/or fraud management system.

The key to this new generation of voice biometrics is its “passive” operation. Calls are silently screened, and no additional questions are asked in order to match voiceprints. Agents proceed normally until they are alerted to any fraudster detection by the system. Today’s passive operation of voice biometrics provides significant advances compared to historical approaches, which required agents to “actively” ask voiceprint verification questions, which can result in customer frustration and longer overall calls.

For higher performance, these systems combine voice biometrics with predictive analytics that factor in other non-biometric variables, including account or call metadata. With these advanced capabilities, organizations can benefit from system accuracy at large call volumes.

Key Benefits for Contact Centers
Solutions focused on helping banks, card issuers, and providers of wealth management services can help deter or prevent fraudulent activities by securely and cost-effectively authenticating customers during interactions on the phone—without negatively impacting the overall customer experience. Considering the following benefits not only helps a financial institution reduce fraud, but also gain visibility into patterns of fraudulent interactions and authenticate legitimate customers quickly.

  1. Stronger fraud protection. Voice biometrics can significantly reduce fraud dollar losses and the number of fraud attacks using voiceprint screening that detects “repeat” calls by known voices, and repeat calls constitute the majority of fraud attacks on the contact center. Organizations can help subvert many of these attacks with today’s advanced biometrics, reducing loss and deterring fraudsters who may then move on to other targets. These benefits make it a rewarding solution for voiceprint tracking.
  2. Increased visibility into fraud. Voice biometric solutions have the capability to track an individual fraudster’s voice even as he/she calls across multiple accounts and time periods. As a result, it can become a rich database and source of information regarding fraudster activities and behaviors. With this intelligence, the contact center can gain important visibility into fraud patterns and details that weren’t available previously via a single source. These insights can be used to further improve contact center security and efficiency.
  3. Reduced vulnerability to social engineering. Although fraudsters can trick or coerce agents, their smooth talk doesn’t fool a voice biometrics engine. By alerting an agent during a call or via the fraud management system even after an agent has been successfully bypassed, advanced voice biometrics solutions can reduce the contact center’s vulnerability to the social engineering of agents.

Tips for Getting Started with Voice Biometrics
As contact centers begin a voice biometrics deployment, it’s important to take several things into consideration. In doing so, organizations will maximize their return on investment (ROI) and speed up the success of a fraud prevention program. The following are recommendations for kicking off a successful fraud and authentication program:

While many contact centers suspect that professional fraudsters are rampant, they can’t always quantify the problem. By conducting a voice biometric analysis across a small sample of fraud calls, an organization can quantify the problem to find eye-opening results. A professional voice biometric analysis can map out and identify any fraudster voiceprints, while also measuring their concentration or repeat rate. If the repeat rate is high (e.g., five fraudsters account for 90% of the total fraud calls), the value of a voice biometric system to detect fraudsters becomes clear.

To be successful, a voice biometrics system must maintain high accuracy, as it relates to fraud detection and false positive rates. Field results show that system accuracy is greatly increased by performing predictive analytics across other factors such as call metadata (e.g., ANI, duration) or account metadata. These factors are taken into consideration along with the voice biometric match—and, as a result, the system’s accuracy can reach extended levels. The addition of predictive analytics is required to achieve high accuracy at scale, as call volumes and fraudster databases become large.

Once your voice biometrics system is in place, make it a regular practice to review fraud activity and system performance—and this can be completed as often as weekly. This type of managed service is especially important during the first 12 months of operation, because it is tied to the following key benefits:

Performance optimization. With multiple parameters to configure in a voice biometric implementation, optimizing performance can significantly improve overall system performance. Regular feedback and tuning over time can help analysts dramatically improve detection rates and false positive rates in the first few months.

Actionable insights from fraud trends. As the voice biometric system detects calls by known fraudsters over a period of time, it also collects enormous data on the individual (i.e., his/her time of activity, churn of phone numbers employed, weakness of certain security questions and more). With access to this information, analysts can regularly inspect the data to gain unique insights into fraud trends and new attack vectors that can benefit the contact center.

As fraudsters evolve with new techniques, so should the contact center in order to protect identities and assets, and help to ensure quality customer experiences. Call screening against a fraudster voiceprint database presents a compelling approach in today’s challenging environments, and it’s worth ensuring that your organization is safeguarded.

– Reprinted with permission from Contact Center Pipeline,

Related Articles

Back to top button