System and method for analyzing and classifying calls without transcription via keyword spotting

A method in a computer system for analyzing a telephone call to assess a classification of the telephone call, the method comprising:

• identifying, by the computer system, a telephone call to monitor;
• determining, by the computer system, a domain or entity associated with the identified telephone call;
• loading, by the computer system, one or more grammars for the determined domain or entity, wherein the one or more grammars contain one or more keywords and associated parameters that are correlated with a likelihood of a particular classification for telephone calls in the domain or entity;
• receiving, by the computer system, an audio signal of the identified telephone call, wherein the audio signal of the telephone call is not transcribed in its entirety;
• using, by the computer system, the one or more loaded grammars to monitor the received audio signal and identify one or more keywords and associated parameters that are present in the received audio signal;
• loading, by the computer system, a statistical model to determine the classification of the identified telephone call;
• using, by the computer system, the loaded statistical model to determine the classification of the identified telephone call based at least on the identified keywords and associated parameters;
• and
• outputting, by the computer system, the determined classification.

A non-transitory computer readable storage medium with instructions stored thereon that, when executed by a computing system, cause the computing system to perform a method for analyzing a telephone call to assess a classification of the telephone call, the method comprising:

• identifying, by the computing system, a telephone call to monitor;
• determining, by the computing system, a domain or entity associated with the identified telephone call;
• loading, by the computing system, one or more grammars for the determined domain or entity, wherein the one or more grammars contain one or more keywords and associated parameters that are correlated with a likelihood of a particular classification for telephone calls in the domain or entity;
• receiving, by the computing system, an audio signal of the identified telephone call, wherein the audio signal of the telephone call is not transcribed in its entirety;
• using, by the computing system, the one or more loaded grammars to monitor the received audio signal and identify one or more keywords and associated parameters that are present in the received audio signal;
• loading, by the computing system, a statistical model to determine the classification of the identified telephone call;
• using, by the computing system, the loaded statistical model to determine the classification of the identified telephone call based at least on the identified keywords and associated parameters;
• and
• outputting, by the computing system, the determined classification.