YOU ARE AT:OpinionReality Check: Voice biometrics and mobile, a perfect match

Reality Check: Voice biometrics and mobile, a perfect match

Editor’s Note: Welcome to our weekly Reality Check column. We’ve gathered a group of visionaries and veterans in the mobile industry to give their insights into the marketplace.

The most user-friendly ways for humans to interact with machines often center on something that every person is born with. One of the best-known recent examples is how smartphones quickly shed the stylus in favor of touch. Today, platforms such as Siri are providing mobile users with the option of simply speaking a text message, a question about nearby merchants or a request for directions. For many users, voice control, or interactive voice response (IVR), offers a faster, better experience than trying to peck away at a physical or virtual qwerty keyboard — not to mention a far safer experience when driving.

The trend toward voice control in the mobile domain creates opportunities to enhance the way that humans interact with companies. A prime example is authentication. Human voices are as individually unique as fingerprints and retinas, so they’re an ideal way for enterprises, government agencies, telcos, banks and other organizations to authenticate employees and customers who want to access information or services remotely.

There are multiple ways to implement voice biometrics, but they all begin by collecting a voiceprint from an authorized user, such as when a policyholder calls her insurance company for the first time. Each subsequent time that she calls, the IVR system can compare her voice with the voiceprint associated with her account to verify her identity.

Although a self-service interaction is an obvious place to implement voice biometrics authentication, it can also be done in the background of a call between a customer and a contact center agent. This latter approach can reduce time spent on manual security questions asked by the agent and significantly enhance security in the contact center by mitigating the threat of social engineering. It can also be used to identify known fraudsters by comparing the caller’s voice with a watch list of known fraudsters’ voiceprints. The biometrics verification process can be completed in as little as 10 seconds, effectively providing a significant real-time layer of security.

In all of these scenarios, the customer benefits because she doesn’t have to type in a passcode — if she even remembers it — or answer a series of questions posed by a live agent before she can finally get the information she seeks. The organization benefits because the streamlined authentication process makes for a better customer experience. The voice biometrics platform also reduces the organization’s customer care costs because each agent now can support more customers and many more calls can be completed successfully within the self-service channel, whether it’s IVR or mobile. Voice biometrics can also enable multifactor authentication, where service providers are struggling to find ways to add a meaningful layer of security beyond the traditional “what you know” layer.

Those benefits have made the business case for large-scale voice biometrics deployments at organizations such as Bell Canada, where more than 2 million customers have signed up for a program that lets them speak a pass phrase to get access to an agent or IVR. Another example is Turkcell Global Bilgi, which handles more than 120 million IVR contacts annually. The company’s agents spent an average of 25 seconds manually authenticating each caller. Voice biometrics reduced that process to aboutfive5 seconds.

Although Turkcell Global Bilgi initially launched this system for a limited number of subscribers, it attracted more attention than expected and reached 2 million users in a short time given that customers like the very simple and fast authentication process of only five seconds.

Today, there are more than 6.5 million registered voiceprints worldwide, according to Opus Research, which expects the number to top 20 million by 2014. “Voice biometric-based solutions are poised to assume the pivotal role of user authentication to support higher levels of trust among users of mobile apps, remote monitoring, distance learning, e-medicine, e-government and a host of other social activities or transactions,” Opus says.

As of December 2011, the Turkcell deployment had more than 4 million enrolled voiceprints, making it the voice biometrics application with the world’s largest number of enrollments.

Leveraging and protecting smartphones

A growing number of voice biometrics deployments have a mobile component because the two technologies complement each other. For example, with smartphone penetration now at or above 50% in countries such as the United States, some organizations are now taking the voice biometrics platform that they originally deployed for IVR and extending it to support their smartphone apps. By providing their employees, customers or both with the option of speaking their authentication information, enterprises can eliminate the frustration and lost productivity that comes with typing on a smartphone’s or tablet’s keyboard.

Voice biometrics also can enhance the security of mobile devices themselves by providing an alternative to entering a passcode to unlock them. That option couldn’t come soon enough because smartphones and tablets used for business increasingly contain confidential information such as files and customer contacts. By providing a user-friendly way to unlock their device, voice biometrics eliminates the security risks that arise when employees try to disable passcode-based locks simply because they find those safeguards a hassle to use.

A well-designed voice biometric system can maintain high accuracy – ensuring both security and a good user experience – even in the face of spotty coverage, interference and other common wireless challenges. That’s because the voice channel is actually not used at all when voice biometrics is used to access a mobile app. Instead, the app can record and then send the passphrase over a secure data connection, bypassing all of the packet loss, transcoding, vocoder limitations and other problems that plague the voice channel.

The ideal voice biometrics solution also should ensure a good user experience by using algorithms that are sophisticated enough to work around problems such as crosstalk and background noise. That accuracy would eventually translate to a better user experience, minimizing the need for callers to repeat themselves.

Enterprises, government agencies and other organizations increasingly realize that customers and employees find voice to be a convenient way to provide authentication information, especially when using a mobile phone or tablet. Unlike passwords and other log-in information, people can’t lose or forget their voice — one more reason why voice biometrics is coming soon to your mobile phone, if it isn’t already there.

Dan Nordale is the vice president of worldwide marketing of Nuance’s enterprise division. In this role, he is responsible for solution marketing, lead generation and field marketing of Nuance’s self-service solutions for the contact center market. Prior to joining Nuance, Dan worked at Genesys for 12 years where he held executive roles in sales, business development and marketing for its high-growth solutions. Dan earned an MBA at Haas School of Business at University of California, Berkeley after completing his undergraduate business studies at University of Alaska, Fairbanks.

ABOUT AUTHOR