Is Voice Authentication Ready for Prime Time?
Many forms of biometric authentication are emerging as potential solutions to banks, including fingerprints, hand geometry, eye retina, facial and voice. Of these, voice holds one distinct advantage over the other forms: the ubiquity of cell phones offers the capability for extending voice biometric authentication on a mass scale. A cursory Google search for “Voice Biometrics” currently yields 50 companies offering related solutions.
However, the mere existence of a technology and delivery method does not mean it addresses business problems cost-effectively. For voice biometric authentication to be successful, it requires efficaciously executing on three elements: meeting customer expectations with a solution that is simple, easy, and convenient; satisfying a real business problem, like reducing fraud losses; and deploying a solution that is cost-effective to implement, operate and sustain.
While banking service providers are planning the infrastructure for voice support, voice solutions companies are competing to provide higher quality and lower costs. The ultimate result depends on if the service can be cost-justified through improved customer acquisition and retention, reduced fraud losses, re-use across the enterprise, and lower service costs from vendor competition.
Banks Take Notice
A recent article in the Wall Street Journal noted that while voice recognition technology has been around for decades, the reliability has improved considerably in the last few years, making it a viable option for financial services companies. Barclays PLC, Vanguard Group and TD Bank Group have already begun to deploy this technology while Irish Bank of Ireland, Banco Santander in Mexico, and United Bank of Pakistan currently have or are planning to soon conduct voice biometric pilots of their own, according to the article.
Minneapolis-based U.S. Bank is dipping its toes into these waters as well, according to the Las Vegas Review-Journal, which quotes Chief Innovation Officer Dominic Venturo as saying, “Passwords are not optimal for mobile devices. They have small screens and keyboards, making it difficult to type in a password. From our perspective, voice biometrics is more secure, while for consumers it’s easier to use.”
Beyond its value as an identity authenticator superior to passwords, voice recognition technology also has the potential of spurring mobile banking to the next level. With voice as a reliable authentication means, financial institutions can address regulatory guidelines (such as those from the FFIEC) while offering many common banking functions with greater security and hands-free convenience. Imagine selecting your bank’s app, touching a microphone icon and saying, “Pay Acme Services one-hundred dollars on Thursday.” The system would verify your voice, schedule your payment and provide an immediate confirmation. Transformational services like this could quickly bring even more voice services into mainstream usage.
And then there’s e-commerce. The retail payments space has received enormous attention from hundreds of companies seeking a claim of the mobile payments ecosystem, with solutions that use near-field communications (NFC) and recently Host Card Emulation (HCE), Quick Response Code (QR) and bar codes, geo-location, etc. However, little has changed for card-not-present payments. The U.S. Census Bureau estimates that retail e-commerce accounts for only 4.7% of total retail sales. Why the holdup? One possibility is the absence of customer-centric options and mechanisms to affordably deploy better authentication solutions on a mass scale.
In 2012, companies reported losing an average of 0.9% of total online revenue to fraud, or roughly $3.5 billion, according to CyberSource. While retail payment innovation is on fire, remote payments have yet to reach room-temperature. Shockingly, the same manually arcane method of simply providing payment card data dominates the e-commerce space as was used for mail and telephone orders 30 years ago.
While extending voice authentication to e-commerce seems like a leap from current plans, consider that customers are increasingly more comfortable with voice recognition, which could easily extend to voice authentication for other functions. Product evolution occurs in incremental steps so that voice migration will likely take a path from customer care authentication and password proxies to voice banking and e-commerce. Finally, customer accessibility to mobile phones and smartphone apps opens a new two-factor channel to currently weak authentication methods previously unavailable.
The cost to process a voice transaction won’t ever be as low as an online or mobile transaction due to the voice component. However, if the service can be shown to improve customer acquisition and retention rates, help direct transaction processing out of higher-cost channels like branches and reduce fraud, the incremental expense may be cost-justified.
Costs can be addressed on other fronts as well. For example, the payments landscape is dotted with new solutions and more competition. The FIDO (Fast IDentity Online) Alliance is a biometrics standard group dedicated to “developing technical specifications that define an open, scalable, interoperable set of mechanisms that reduce the reliance on passwords to authenticate users.” Such standards are intended to remove technical barriers and provide a level-playing field to build scale and deploy at lower costs.
Further, as banks deploy voice biometrics for customer authentications, the registered user base can be extended to other authentication needs such as step-up authentication, stronger out-of-band authentication and faster and easier person-to-person (P2P) transfers. Reusability of the solution across services should provide greater value to the bank and, with economies of scale, drive down the cost of the solution.