Several weeks ago we published a blog post entitled “When to Use Speech Recognition.” In that post, we provided some examples of when it makes more sense to use touch tone input over speech recognition. In most cases, there is far less user frustration to simply ask callers to enter information via their keypad. Collecting information in this fashion also insures a higher degree of accuracy and lower opt out rates. So, when does it make sense to use speech recognition in favor of touch-tone input?
We’ve all called super slick IVR applications that heavily rely on speech recognition to collect information. One good example of such an IVR application is the FedEx package tracking system. It’s logical for companies the size of FedEx to invest heavily in a speech application of this type. Since this IVR processes millions of calls a day, there is clear ROI on such a large investment of time and money to train speech engines to accurately detect a caller’s utterance. In addition, the general appeal of speech adds value to their brand.
But when should a small or medium size business use speech recognition? We feel the best use of speech recognition in lower call volume applications is to collect data that cannot be entered via a caller’s keypad. Of course, to see ROI on such an investment, a company would have the need to collect data such as first and last name or a customer’s address during their most frequent calls. Some examples for speech recognition grammars that perform these types of functions can be found here: Speech Recognition Grammars
Leave a Reply