ASR - Automatic Speech Recognition

ASR is Automatic Speech Recognition - technology that allows computers to interpret human speech.

Simple ASR can be used to replace number selection menus so that callers can use prescribed words to ask for what they want – such as ‘sales’, ‘accounts’ or ‘complaints’.

Advanced versions of ASR use Natural Language (NL) which lets people use their own words to have a conversation with a machine. They are already in use with Siri and Alexa but are finding greater use in contact centres to provide self-service engagement for customers. Natural Language used in an IVR can be used to gather information, verify callers, answer question or route calls for instance.

Advanced ASR – using NLP – can be programmed to ‘learn’ from each interaction.

Human tuning: This involves human programmers going through the conversation logs of the ASR software interface and looking at the commonly used words that it had to hear but which it doesn’t have in its pre-programmed vocabulary. There words are then added to the software so that it can expand its comprehension of speech.

Active learning: This involves the software itself being programmed to independently learn, retain and adopt new words to constantly expand its vocabulary and embrace new ways of customers speaking and saying things.

Read more about speech recognition IVR