Configuring Automatic Speech Recognition

IVR Routing supports integration with an Automatic Speech Recognition server to provide Automatic Speech Recognition capabilities for collecting spoken caller input. Automatic Speech Recognition enables callers in IVR Routing workflows to provide information verbally rather than by entering digits or text using a phone keypad.

Automatic Speech Recognition can be added to a site and its media servers through associating an Automatic Speech Recognition server to the site. For information on supported Automatic Speech Recognition servers, see the MiContact Center Business and MiVoice Analytics System Engineering GuideMiContact Center Business System Engineering Guide.

Activities using Automatic Speech Recognition require a Confidence Threshold for speech input. The Confidence Threshold determines the threshold for accepting speech input and routing down the appropriate branch of the activity. A high Confidence Threshold improves the accuracy of the input while increasing the likelihood that speech input will be considered invalid. A low Confidence Threshold lowers the chance that speech input will be considered invalid, at the cost of accuracy. The default Confidence Threshold for activities using Automatic Speech Recognition is 50%. It is the responsibility of contact centers to determine the acceptable Confidence Threshold for speech input in their workflows.

IVR Routing Automatic Speech Recognition works with IVR Routing Text-to-Speech. For more information on Text-to-Speech, see "Configuring a Text-to-Speech server".