
IVR systems can handle the dissemination of information from a large and complex database by using text to speech technology, which converts ASCII text to computer generated speech. This is often useful for speaking out names of individuals, streets, cities, companies, or products. The conversion is very accurate and the computer voice has mostly lost its mechanical nature and sounds quite human.
The input of data to IVR systems can be made more user friendly by using speech
recognition technology. In its simplest form speech recognition in the typical
IVR application where caller populations are large and diverse is limited to
recognizing a small vocabulary consisting of simple digit strings and words
such as yes, no, pound and star. This level of speech recognition is used to
replace the necessity of using the telephone keypad. However speech recognition
technology continues to improve to accept broader vocabularies and streamline
the call flow. For instance if an application requires input of information
regarding various auto service that has been performed, the traditional approach
would be:
For engine, press 1
For electrical system, press 2
For body work, press 3
For shocks and wheels, press 4
With our advanced speech recognition the caller would simply say "shocks and wheels". Virtually every application can benefit to some degree from the use of speech recognition.