On the Verge of Voice Commerce
After obtaining 401(k) information from Manulife (Toronto), a customer once marveled, "She understood me, even without my teeth." The "she" the customer was referring to wasn't a she at all, according to Kendall Kay, chief strategy officer, Manulife, but a synthesized speech system in use at the insurance company. A synthesized voice that once seemed robotic can now sound so natural that customers, thinking it a live person, try engaging the technology in conversations about the weather, Kay notes.
Speech recognition-the flip side of voice technology-is becoming equally advanced, with systems that can perform sophisticated functions, such as parsing sentences, explains Howard Sachar, global manager, insurance and banking research center, IBM (Armonk, NY).
Voice commerce, which Sachar defines as doing and enabling commerce using speech for either input or output, seems like the natural progression for insurers currently selling to and servicing their customers online. A company's success with voice commerce is dependent on a number of factors. According to Sachar, an insurance company thinking about implementing a voice technology must ask itself, "What am I going to be giving access to, how can I get to that access and what do I need to do in terms of protecting that asset and caring for it? Insurers also must analyze the extent to which customers will be tolerant of a voice commerce application."
There are many options for companies exploring the use of voice for commerce. Beyond the interactive voice response (IVR) system, which runs touch-tone applications, there is speech recognition that can identify single spoken words. More complex recognition systems can identify a single word within a string of words. A customer calling an insurance company, for example, could say, "I'd like to file a claim." Then the system would prompt the client through the entire process, according to Steve Ehrlich, vice president of marketing, Nuance (Menlo Park, CA), a natural voice interface software provider.
Beyond merely recognizing words lies a voice system that not only can identify what is being said, but who is actually saying it. Biometric voice printing, a sophisticated method of voice recognition, can be used instead of a PIN. A voiceprint, like a fingerprint, is unique to each individual. "There is a lot of value going with biometric voice print," says Bob Zurek, senior analyst, Forrester Research (Cambridge, MA). "Such a system might ask you to speak, and then your voice is recorded and stored in a database. The system also stores a challenging voice and the two are compared each time you enter the system, so it can positively identify you."
Conrad Rossow, director of emerging technologies, Minnesota Life (St. Paul, $18.4 billion), points out that voiceprints can also shorten the path of movement through the hierarchy of questions for which a system may prompt. "Instead of having to say your name and that you want your 401(k) account information and give your Social Security number, it will automatically know who you are, since it can positively identify you by your voice pattern, " says Rossow.
Tweet This