Lymba Catalyst
Lymba Catalyst is a semantic categorizer product designed to automatically map a natural language utterance into one or more predetermined categories.
Building an interactive voice response (IVR) system currently requires months of expensive manual work. Once the menu choices are determined, testers must make thousands of calls to the prototype system. Their conversations are manually transcribed, and some of the testers' utterances are added to a grammar (list of utterances) that the final system will recognize. The testers then call the system again, and the grammar is gradually tuned to recognize the utterances that callers are most likely to say.
Lymba Catalyst eliminates the need to develop grammars for a directed dialog application, and it removes the costly tuning phase when deploying a dialog application to a customer. It also does not require any data to be gathered for domain specific training, although it can benefit from grammars and domain specific information in Jaguar ontologies.
To learn more about how your organization can benefit from using natural language processing solutions from Lymba, please contact us.