Voice Biometrics Can Secure Hassle-Free Caller Verification

Article Featured Image

Protecting customers and preventing fraud is a top priority for most contact centers, particularly those in financial services, insurance, healthcare, and other sectors involving sensitive and personal customer information. To reduce the risk of exposing that information to fraudsters, companies dedicate millions of dollars and a great deal of time and effort to identifying and verifying their customers.

Unfortunately, the time and cost of identification and verification is increasing while the benefits are dropping. The verification processes used in many contact centers are far too open to fraud, as the answers to many security questions can be social-engineered with little effort. When a company tries to use more secure questions, its customers will often complain that the process is cumbersome and the questions difficult to answer. Changes and improvements are needed, but few organizations have come up with a good solution to this dilemma.

Enter Voice Biometrics

Voice biometrics has great potential to help companies and consumers minimize both the risk of fraud and the time spent on identity verification. DMG Consulting defines voice biometrics as “technology that applies algorithmic formulas to the features and physical characteristics of an individual’s voice to create a unique, non-replicable voiceprint for the purpose of identification and authentication when compared to a live audio stream.” Voice biometrics technology uses the distinct characteristics of speakers’ voices to identify them. This includes voice features such as language, accent, pitch, tone, etc., along with physical characteristics such as the shape of the larynx, lips, tongue, mouth, teeth, and so on.

Using live audio streams or recorded interactions, the technology extracts the speaker’s unique voice features and applies algorithms to create and “enroll” the voiceprint or voice signature in the system. Depending on the method used by a voice biometrics solution, voiceprints can be created based on specific words or phrases the speaker is prompted to utter (text-dependent), a free-form conversation (text-independent), or both. Regardless, the resulting voiceprint is a mathematical representation of voice features; it is not an image or a recording, and it cannot be played, listened to, replicated, reverse-engineered, or used outside of the system in which it was created.

A Technology Made for the Contact Center

The primary purpose of voice biometrics in the contact center is to automate the verification process so that once a voiceprint is set up, callers do not need to answer a number of security questions—which, ironically, often contain the same sensitive information they are trying to protect. Instead, the voice biometrics engine compares the voice features of the live audio stream to the enrolled voiceprint. The live audio is scored against user-defined thresholds to determine whether the live audio is a match or mismatch, or whether it requires further verification by alternate means. As each person’s voice has unique characteristics that are essentially impossible to replicate, voice biometrics is considered the most accurate and least intrusive way of authenticating callers.

Impediments to Adoption

Given voice biometrics’ major contributions, benefits, and savings for consumers, businesses, and society in general, it’s surprising that adoption of this 20-plus-year-old technology has been sluggish at best. In the fourth quarter of 2016, DMG Consulting conducted a worldwide benchmark study on consumer channel preferences, asking consumers how they felt about companies using voice biometrics to verify identity. The answers to this survey question explain why it’s taken so long for voice biometrics to be broadly accepted in the market. A significant majority of consumers, 60.4 percent, are either unfamiliar (“not sure/don’t know”) with voice biometrics or are uncomfortable with it; only 39.6 percent are comfortable with the idea of using this technology.

Acceptance of Voice Biometrics

Biometrics in general, and voice biometrics in contact centers and websites, can reduce the risk of fraud, saving consumers and companies millions in losses and operating expenses. Recent innovations in voice biometrics make this technology viable, and the entrance of new competitors is making it more cost-effective. DMG’s benchmark study findings show that lack of knowledge and understanding about this technology is impeding its adoption—which is a relatively easy issue to solve. It’s time, then, for vendors to invest in educating the market about the value proposition and benefits of voice biometrics.

Donna Fluss (donna.fluss@dmgconsult.com) is the president of DMG Consulting, a provider of contact center, analytics, and back-office market research and consulting.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues
Related Articles

Speech Analytics and AI Is a Winning Combination

Service, quality management, and the customer journey will all see big gains

11 Keys to Designing Effective IVAs

Businesses and customers alike can benefit from quick, personalized, intelligent self-service.