Like Text to Speech, Speech Recognition is a major technology that can be used to strengthen and add flexibility to a wide range of IVR services and campaigns. The BCH IVR supports speech recognition engines from DQ Labs and Scansoft OSR.

Grammars are used by the engines to define words and phrases that can be recognised. These grammars can be predefined and loaded when a system is started up, or alternately created dynamically to recognise a wide range of words on the service running for example, a list of names loaded from a database.

The grammars use the open standard voice XML, making grammars easy to design and deploy on a running IVR system. Grammars can be defined that recognise a large number of individual words, or even words from phrases. For example - 'Please take me to the airport' could recognise 'take' and 'airport', so that the phrase, 'I would like someone to take me to the airport' could also be spoken by the caller with the same effect.

Sophisticated services can use Speech and Text and Speech together. For example, an address recognition system that uses Speech Recognition to understand an address could then use 'Text to Speech' to play it back to the caller confirming the details, therefore using the two technologies in conjunction with each other provides even greater flexibility and functionality.

Related Products