Biometrics Help Virtual Assistants Help Customers Faster

Speech and natural language technology vendor Interactions in late November added voice biometrics to its Intelligent Virtual Assistant, allowing the application to seamlessly identify and authenticate callers using the

unique characteristics of their individual voices. Fellow speech technology vendor Nuance Communications followed that up a few weeks later with its mid-December release of Nina ID 2.0, which adds integrated multifactor authentication to its own Nina Virtual Assistant for customer service.

“Voice biometrics has been around for a while, but there was a huge uptick in it last year,” says Mary McKenna, Interactions’ director of product management.

Part of the reason for that, McKenna explains, is that the technology is far superior today compared to what was available just a short time ago. While voice enrollment (the process of registering the initial voiceprint upon which all subsequent utterances are measured) typically took around 40 seconds, and the actual authorization took about 20 seconds, Interactions Voice Biometrics cuts those figures down to single digits. And, unlike solutions of the past, the solution doesn’t require call center agent intervention for enrollment, she adds.

Likewise, with the introduction of Nina ID 2.0, customers can be automatically authenticated by saying simple voice passphrases or taking selfies. Nina uses AI-powered voice biometrics and face recognition to confirm the identity of users. Once consumers create their own unique voiceprints or face prints, authentication can take place in a matter of seconds.

With the Nina ID risk engine, authentication factors that are provisioned by third parties, such as fingerprint recognition on smartphones, can also be managed to ensure that desired security levels are maintained regardless of the authentication method.

With modern biometrics solutions, it’s also possible to offer multiple identification modalities in a single application. Interactions’ latest offering, for example, offers the industry-standard dual verification, but also, if desired, triple verification, McKenna says. Dual verification involves voice coupled with an account ID or PIN. Three-factor verification involves voice, account ID or PIN, and automated verification of the device from which one is calling.

Users can choose active or passive authentication. Active authentication matches a caller’s unique voiceprint to a passphrase that is predetermined or generated at the time of the call. With passive authentication, the voiceprint is matched as the caller speaks naturally; no special phrase or password is required.

Cloud deployment options are also helping to speed adoption of voice biometrics in virtual assistants, according to McKenna. “It’s a very simple implementation with our voice virtual assistant. Since this is a hosted solution, there is no hardware to buy, storage management, maintenance, or tech people to bring in. It is very easy to turn on.”

Dan Miller, founder and lead analyst at Opus Research, isn’t surprised by the trend. “Intelligent authentication goes hand in hand with the expansion of virtual agents and intelligent assistants,” he says.

That expansion has been a long time coming. Introduced in 2012, Nina ID originally offered integrated voice biometrics and quickly expanded to support virtual assistant–powered customer service on websites, within messaging apps, and via text message.

“Consumers today need to be able to reach out and engage with brands quickly and easily, wherever they are, and that’s often from a mobile device,” said Robert Weideman, general manager and executive vice president of the Nuance Enterprise Division, in a statement.

For businesses, McKenna foresees wide implementation possibilities across financial services, insurance, telecommunications, government, healthcare, e-commerce, and other industries that have high call volumes requiring customer authorization.

Approximately 30 percent to 50 percent of all fraud incidents are initiated by phone, according to Interactions. The financial services industry alone reported more than 1 million incidents of fraud in the first half of 2016, equating to one incident every 15 seconds, a 53 percent increase compared to the same period last year.

McKenna adds that gaining the benefits from voice biometrics will rely on companies successfully enrolling high percentages of their customers in voice verification. Enrollment success will depend largely on marketing and how companies set up the enrollment process, McKenna says, adding that in trials some companies have had enrollment rates as high as 90 percent with some voice biometrics solutions.