Call flow object model in a speech recognition system
View Patent ↗Systems and methods are provided for embodiments of a speech recognition system call flow object model. The systems and methods organize and execute, for example, multiple question directed dialogs, overview dialogs, or natural language directed dialogs. In certain embodiments, the organizing and executing of the natural language directed dialogs uses primary and secondary concepts without requiring a structured response. The systems and methods enable the call flow designers to define particular call flows without requiring the designers to perform any programming or coding.
1. A speech recognition system implemented on a computer, the system comprising:
a speech recognition engine comprising a plurality of modules for implementing a call flow having directed dialogs that a caller moves through during a call with the system, the speech recognition engine comprising
a global module having defined therein primary concepts, and secondary concepts associated with the primary concepts, that are available to the caller throughout the call flow;
a module configured to record a natural language audio input response from a caller; and
a module configured to process at least a portion of the call flow and recognize one or more primary concepts and at least one secondary concept associated with each of the one or more primary concepts in an audio input comprising natural language, apply a score to each primary and secondary concept pair found in the audio input, and determine the caller response based on the concept pair with the highest score.
2. The speech recognition system of claim 1 , wherein the audio input comprises a non-structured response.
3. The speech recognition system of claim 1 , wherein the call flow executes a response chain based on the audio input.
4. The speech recognition system of claim 3 , wherein the score of the primary and secondary concept pair identifies the response chain to execute.
5. The speech recognition system of claim 1 , wherein the words are found within vocabularies, the vocabularies found within grammars, and the grammars found within at least one or more of the modules.
6. A non-transitory computer-readable medium comprising software configured to perform the steps of:
transmitting a prompt to a caller;
receiving a natural language audio input from the caller in response to the prompt;
identifying a first word, associated with a primary concept, and a second word, associated with a second concept, in the audio input;
generating a score based on the primary and secondary concepts; and
executing a response based on the score.
7. A method of defining a call flow in a speech recognition system implemented on a computer, the speech recognition system having a speech recognition engine including a plurality of modules for implementing a call flow having directed dialogs that a caller moves through during a call with the system, the method comprising:
defining in the speech recognition engine
a global module having defined therein primary concepts, and secondary concepts associated with the primary concepts, that are available to the caller throughout the call flow;
a module configured to record a natural language audio input response from a caller; and
a module configured to process at least a portion of the call flow and recognize one or more primary concepts and at least one secondary concept associated with each of the one or more primary concepts in an audio input comprising natural language, apply a score to each primary and secondary concept pair found in the audio input, and determine the caller response based on the concept pair with the highest score.
8. The speech recognition system of claim 7 , wherein the audio input comprises a non-structured response.
9. The speech recognition system of claim 7 , wherein the call flow executes a response chain based on the audio input.
10. The speech recognition system of claim 9 , wherein the score of the primary and secondary concept pair identifies the response chain to execute.
11. The speech recognition system of claim 7 , wherein the words are found within vocabularies, the vocabularies found within grammars, and the grammars found within at least one or more of the modules.
12. A method of processing an audio input for speech recognition, the method comprising:
transmitting a prompt to a caller;
receiving a natural language audio input from the caller in response to the prompt;
identifying a first word, associated with a primary concept, and a second word, associated with a second concept, in the audio input;
generating a score based on the primary and secondary concepts; and
executing a response based on the score.
13. The speech recognition system of claim 12 , wherein the words are found within vocabularies, the vocabularies found within grammars.