System and Method for Determining the Compliance of Agent Scripts

• US 20160329049A1
• Filed: 07/22/2016
• Published: 11/10/2016
• Est. Priority Date: 08/28/2013
• Status: Active Grant

• Current Assignee: Verint Systems Incorporated
• Original Assignee: Verint Systems Limited (Verint Systems Incorporated)
• Inventors: Iannone, Jeffery Michael, Wein, Ron, Ziv, Omer
• Granted Patent: US 10,573,297 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G10L 15/04 Segmentation; Word boundary detection
• CPC Class Codes: G10L 15/06 Creation of reference templates; Training of speech recognition systems, e.g. adaptation to the characteristics of the speaker's voice G10L15/14 takes precedence
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/10 using distance or distortion measures between unknown speech and reference templates
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/0635 updating or merging of old and new templates; Mean values; Weighting
• CPC Class Codes: G10L 2015/088 Word spotting

Abstract

Systems and methods of script identification in audio data obtained from audio data. The audio data is segmented into a plurality of utterances. A script model representative of a script text is obtained. The plurality of utterances are decoded with the script model. A determination is made if the script text occurred in the audio data.

Claims

1. A method of script identification in audio data, the method comprising:

obtaining audio data;

segmenting the audio data into a plurality of utterances;

obtaining a script model representative of a script text;

decoding the plurality of utterances by applying the script model to the plurality of utterances;

determining if the script text occurred in the audio data from the decoded plurality of utterances;

applying speech analytics to the plurality of utterances to identify at least one event;

upon identification of the at least one event selecting the script model based upon the identified at least one event; and

decoding a subset of the plurality of utterances occurring after the identified at least one event with the selected script model,

wherein decoding the plurality of utterances comprises applying a plurality of script models to the plurality of utterances and producing an indication of script models identified in the plurality of data.

3. The method of claim 1, further comprising:

filtering the plurality of utterances to include only utterances attributed to a customer service agent;

extracting acoustic features from the filtered plurality of utterances; and

using the extracted acoustic features in decoding the plurality of utterances.

5. The method of claim 1, wherein if the script text is determined to have occurred in the audio data, further comprising:

transcribing the utterance containing the script to produce an utterance transcription;

comparing the script text to the utterance transcription;

determining a script accuracy.

6. The method of claim 5, further comprising evaluating a compliance of the audio data with a script requirement threshold by comparing the determined script accuracy to the script requirement threshold.

7. The method of claim 6, wherein the script requirement threshold is a minimum word error rate.

8. The method of claim 6, wherein if the audio data is evaluated as non-compliant, further comprising:

initiating at least one remedial action wherein the at least one remedial action is selected from operating a graphical display to present on screen guidance to a customer service agent, presenting additional information to a customer;

producing an alert of a non-compliant script.

9. The method of claim 5, wherein if the script text is determined to have not occurred, initiating at least one remedial action.

10. The method of claim 1, further comprising:

compiling the script model from at least the script text; and

modifying the script model to include acceptable variations in the script text.

11. The method of claim 10, wherein the script model is further compiled from at least one speaker acoustic model.

12. The method of claim 1, wherein the audio data is streaming audio data.

2. (canceled)

4. (canceled)

13-20. -20. (canceled)

21. A method of script identification in audio data, the method comprising:

obtaining audio data;

segmenting the audio data into a plurality of utterances;

obtaining a script model representative of a script text;

decoding the plurality of utterances by applying the script model to the plurality of utterances;

determining if the script text occurred in the audio data from the decoded plurality of utterances;

applying speech analytics to the plurality of utterances to identify at least one event;

upon identification of the at least one event selecting the script model based upon the identified at least one event;

decoding a subset of the plurality of utterances occurring after the identified at least one event with the selected script model;

transcribing the utterance containing the script to produce an utterance transcription;

comparing the script text to the utterance transcription; and

determining a script accuracy.

22. The method of claim 21, further comprising evaluating a compliance of the audio data with a script requirement threshold by comparing the determined script accuracy to the script requirement threshold.

23. The method of claim 22, wherein the script requirement threshold is a minimum word error rate.

24. The method of claim 22, wherein if the audio data is evaluated as non-compliant, further comprising:

initiating at least one remedial action wherein the at least one remedial action is selected from operating a graphical display to present on screen guidance to a customer service agent, presenting additional information to a customer;

producing an alert of a non-compliant script.

25. The method of claim 21, wherein if the script text is determined to have not occurred, initiating at least one remedial action.

26. The method of claim 21, wherein the audio data is streaming audio data.

27. A method of script identification in audio data, the method comprising:

obtaining audio data;

segmenting the audio data into a plurality of utterances;

obtaining a script model representative of a script text;

decoding the plurality of utterances by applying the script model to the plurality of utterances;

determining if the script text occurred in the audio data from the decoded plurality of utterances;

compiling the script model from at least the script text; and

modifying the script model to include acceptable variations in the script text.

28. The method of claim 27, wherein the script model is further compiled from at least one speaker acoustic model.