Menu-based, speech actuated system with speak-ahead capability
View Patent ↗An interactive voice response system has speak-ahead capabilities similar to type-ahead IVR systems by determining multi-level grammars for responses. Preferably, an existing IVR application is processed automatically to generate a multi-level grammar database that can then be used in recognizing multi-level responses by a user.
1. A method of creating a grammar used in an interactive voice response system generating prompts based on a menu, comprising:
obtaining a first grammar specifying a first set of responses to a first prompt;
obtaining a second grammar specifying a second set of responses to a second prompt subsequent to the first prompt;
automatically producing a third grammar with a third set of responses to the first prompt that includes the first set of responses and concatenated responses, each combining one response in the first set of responses and one response in the second set of responses to form a sequence of the one response in the first set of responses and one response in the second set of responses.
2. An interactive voice response system, comprising:
a storage unit storing data defining prompts to users, grammars of acceptable responses and a menu of the prompts, all ready for use in said interactive voice response system;
a composite grammar generator, coupled to said storage unit, automatically generating a composite grammar of acceptable responses by:
(a) obtaining a first grammar specifying a first set of responses to a first prompt;
(b) obtaining a second grammar specifying a second set of responses to a second prompt subsequent to the first prompt; and
(c) automatically producing a third grammar with a third set of responses to the first prompt that includes the first set of responses and concatenated responses, each combining one response in the first set of responses and one response in the second set of responses to form a sequence of the one response in the first set of responses and one response in the second set of responses; and
a speech recognizer, coupled to said storage and said composite grammar generator, recognizing speech input in response to a current prompt, as the sequence of the one response in the first set of responses and one response in the second set of responses, based on the composite grammar of acceptable responses that includes an utterance responding to the subsequent prompt.