Banking on speech to prevent fraud.
And as momentum grows for replacing PINs, passwords, and security questions as a means to authenticate customers, the traditional voice biometrics markets have had to undergo a lot of changes. The results have been more passive systems and ones that rely on more than one modality to separate customers from fraudsters.
When banks rely on PINs, passwords, or security questions to verify callers' identities, 85 percent of customers find the approach to be a nuisance and have abandoned the process as a result, according to Opus Research. Banks are coming to that conclusion as well.
"We heard from our customers that verifying their identity through answering security questions was a frustrating process, and we listened and made meaningful changes," Bob Rivers, president and chief operating office at Boston-based Eastern Bank, said in a statement following his firm's adoption of passive voice biometrics technology from Nuance Communications last year. "Voice biometrics allows our customers to get service in the most natural and intuitive way, through their voice. The end result is a better and more effortless experience for our customers."
Eastern Bank customers can enroll their voiceprints into an authentication system that passively checks for a successful match on subsequent calls. Once enrolled, callers engage in natural conversation with customer service agents while the system compares their voiceprints in the background.
English firm Barclays has also implemented a similar passive voice authentication system for its high-value customers. And in Canada, RBC has also made passive voice-based authentication, supported by Nuance Communications' FreeSpeech technology, available to its 18 million clients. RBC piloted the technology in early 2015 and found that it was well-liked by callers. In fact, more than 90 percent of RBC's customers who were given the opportunity to opt in for the service have elected to do so.
Manulife, another Canadian financial services firm, also offers a voice-based authentication service based on Nuance's VocalPassword technology. The new system, which it launched in the summer, relies on the passphrase "At Manulife my voice is my password" or its French equivalent, "A Manuvie, mon mot de passe, c'est ma voix."
Dutch banking firm ING's approach incorporates both voice biometrics and multifactor authentication that enables customers to initiate payments on their mobile phones using voice, face, and fingerprint recognition. The authentication strategy, which also includes technology from Nuance, will allow ING customers to make bank transfers and check balances through a natural, conversational interface.
"Thanks in part to the inclusion of biometric applications, we are able to make banking faster, smoother, and easier for our customers and improve access," Jeroen Losekoot, Internet and mobile marketing manager at ING, said in a statement. "This allows them to have greater control over their finances."
Wells Fargo expects the same results from its deployment of SpeechPro's VoiceKey.OnePass voice biometric algorithms and supplemental technologies. With VoiceKey.OnePass, business customers can log in to Wells Fargo's mobile apps using face and voice authentication. The solution's accuracy derives from a combination of voice authentication, facial recognition, and anti-spoofing liveness tests.
"We believe this is a turning point for the industry, as this is the first time an authentication system that simultaneously collects voice and face biometrics with additional liveness detection features has been used by customers of a major financial institution," says Alexey Khitrov, president of SpeechPro. "Accuracy of multimodal biometrics finally met the convenience of a one-step process for the user." Sensory, through its TrulySecure technology, also offers multimodal face and voice biometric authentication software. Sensory first launched TrulySecure in July 2014 and updated the solution to version 2.0 in the summer with neural networking and machine learning.
The result of that combination, according to Todd Mozer, CEO of Sensory, is "a seamless user experience" and "virtually impenetrable security against hackers and identity theft without sacrificing convenience or adding undue system costs."
In tests leveraging real-world data collected from thousands of actual users operating a variety of devices in a variety of environments, TrulySecure 2.0 matched the correct user in more than 98 percent of transactions while keeping out imposters 99.999 percent of the time.
Verint Systems has gone even farther with its release of Verint Identity Authentication and Fraud Detection, which combines voice biometrics, speech analytics, pattern matching, and other fraud detection technologies.
Verint Identity Authentication and Fraud Detection is more than a tactical solution. The solution is embedded in the company's broader customer engagement optimization portfolio. It can be added to existing Verint call recording environments or be installed separately.
As part of the portfolio, it also features unification and a shared workflow with Verint Desktop and Process Analytics to provide the relevant guidance to employees as part of their engagements. Depending on where employees are in the customer engagement process, desktop and process analytics can provide specific next-best-action guidance based on identity matching scores. By leveraging other solutions, such as Verint Speech Analytics, organizations can detect, analyze, and act on trends, opportunities, and anomalies.
"You want to have specific guidance in place so that the agent can take the appropriate corrective action when a fraudster is detected," says Steve Williams, vice president and global practice leader for identity analytics at Verint.--Speech Technology staff
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Market Spotlight: Financial Services|
|Publication:||Speech Technology Magazine|
|Date:||Mar 22, 2016|
|Previous Article:||Microsoft opens up its speech APIs to developers.|
|Next Article:||Over heard under heard: under-the-radar speech news.|