TradeHarbor, which supplies voice authentication solutions in finance, health and other sectors, is offering its SaaS (software-as-a-service) solution for integration and applications in cable customer service and TV Everywhere at no initial cost to cable operators. “We are so confident of the significant return on investment for using the Voice Signature Service in the cable industry that we are offering free integration support and free use of the service to prove the ROI,” says former Charter Communications chairman Barry Babcock, who is member of TradeHarbor’s board.
The technology has caught the attention of the studios as well. “The motion picture industry has been looking for a way to implement secure parental controls and remote access of high value content by authentication of the user for nearly a decade,” notes Brad Hunt, former CTO of the Motion Pictures Association of America, who maintains consulting ties to the film industry as president of Digital Media Directions. “TradeHarbor’s VSS (Voice Signature Service) can enable these services while preserving the privacy of the person.” Now, after seeing the technology in action at the CableLabs meeting, Hunt is advising TradeHarbor as well.
The implications of bringing voice recognition into media as a means of authenticating the identity of individuals rather than relying on traditional account verification through passwords and shared secrets could be significant as players in this space seek to make authentication more user friendly with greater personalization of individual usage on household accounts. The key to success is the extent to which the technology has arrived as a truly reliable supplement or even replacement to traditional methods.
Paul Heirendt, president and CEO of TradeHarbor, points to the firm’s track record in other industries as proof that the technology has arrived. “Trust is at the heart of what we do,” Heirendt says.
TradeHarbor’s VSS has been in commercial service since December 2005 when a Fortune 100 healthcare insurer began using the service to enable customers to sign insurance applications over the telephone, Heirendt notes. It’s also used in other insurance programs, finance and various government programs, including a court authentication application in the federal justice system and military applications, he adds.
In May last year customer relationships management supplier Convergys announced it had implemented the TradeHarbor solution across its industrial client base spanning automotive, communications, finance, health care, utilities, transportation, retail and other sectors. The firm’s Voice Authentication Deployment Framework offers a model as to how the technology can be flexibly deployed to service multiple requirements across the enterprise, notes Paul Watson, general manager of on-demand solutions at Convergys.
“Our clients have been asking us for increased authentication security for their customer self-service – for both speech and touchtone applications as well as agent-assisted transactions, but without the traditional capital expenditure hurdle,” Watson says. The framework includes an operations and administrator interface for managing deployments, adjusting security thresholds and generating reports. In addition, any company can easily access Convergys On-Demand Voice Authentication via an API (application program interface) or Web services from an existing IVR (interactive voice response)/Voice Portal platform, he adds.
“Through Convergys we’re already working with six MSOs in customer service enablement areas,” Heirendt says.
As explained by Heirendt, the voice authentication process provides a legally defensible means by which users confirm agreements over the phone by stating a prescribed phrase that serves as their voice signature on the agreement. Their recorded signature later provides the basis for authenticating them when they contact the provider.
“On enrollment, the user is asked to repeat four prompt phrases, which takes about 20 seconds,” he says. “Subsequently, for authentication the user is asked to repeat two phrases, which takes about 9 seconds.”
The phrases the user is asked to repeat are selected by the VSS from an ever-expanding lexicon of identity rich phrases. “The user has nothing to remember or that could be recorded ahead of time by a frauder,” he notes.
The ROI selling point has to do with the time and effort saved in customer service through use of voice recognition as well as the impact it has on cutting losses due to fraud. “Credit card information and a password are very fraudible,” he says. “The current move to best practices in financial services and retail is toward shared secrets, but that takes a lot of time and it often makes consumers uncomfortable. And the idea of a shared secret is an oxymoron. Once it’s shared it’s not a secret. And if a legitimate user gets the secret wrong, it’s really inconvenient.”
The existing authentication process takes 45 seconds to two minutes, he says, adding, “We reduce the time to nine seconds. Even if all you use is the user’s telephone number and the last four digits of their social security number to identify them, it takes 30 seconds versus nine seconds. Even that differential is a significant metric when it comes to delivering ROI on voice authentication.”
Voice authentication isn’t perfect, of course, given the fact that factors like a cold or too much alcohol consumption can distort voices. But TradeHarbor’s proprietary technology includes methods by which such distortions are mitigated, resulting in an error rate of below two percent, Heirendt says.
The VSS provides decision support for assessing risk on a transaction-by-transaction basis by returning a normalized Confidence Score, he adds. The risk balance goes between false accepts and false rejects so that, if a provider wants to maximize customer satisfaction, they can tune the system to err on the side of acceptance over rejection. Where the emphasis is on very high security, the system can be tuned in the other direction.
Were voice signatures to be used to authenticate users in TV Everywhere types of scenarios, the service provider would be able to address specific applications, such as parental controls, or specific ads to individuals, Heirendt notes. “The majority of our customers initially calculate ROI on the basis of eliminating fraud and saving time,” he says, “but then it quickly goes to re-engineering services to exploit the capabilities in other ways.”
But using voice authentication in PC-based TVE access would require that users either call in to a number or that their PCs be equipped with special plug-ins to accommodate direct voice communications with the system over the broadband link. “The experience over basic HTML won’t be satisfactory on the PC,” Heirendt says. “We’re not there yet with normal HTTP delivery, even if you’re using [a CDN provider like] Akamai.”
Accessing TVE via mobile phones and tablets, however, is an entirely different story.
“We just released the ability to interact with specific devices, including iPhones and Android phones,” he says. “The app pops up on the phone. We provide the software code for people who have mobile apps so that when they add our code the application includes voice authentication.” The same can be done with apps running on OS-optimized tablets, he adds.
Eventually as TVE becomes a cloud-based service serving multiple types of devices, direct voice input can become a natural extension of the service, resulting in a far more user-friendly and user-specific applications environment, Heinrendt says. If he’s right, the technology could prove effective well beyond the traditional ROI considerations.