Script compliance and quality assurance using speech recognition
Apparatus and methods are provided for using automatic speech recognition to analyze a voice interaction and verify compliance of an agent reading a script to a client during the voice interaction. In one aspect of the invention, a communications system includes a user interface, a communications network, and a call center having an automatic speech recognition component. In other aspects of the invention, a script compliance method includes the steps of conducting a voice interaction between an agent and a client and evaluating the voice interaction with an automatic speech recognition component adapted to analyze the voice interaction and determine whether the agent has adequately followed the script. In yet still further aspects of the invention, a voice record of the given interaction can be analyzed at least in part to determine whether the agent spoke any extraneous material not dictated by the script, apart from or in addition to ensuring that the voice record contains all material required by the script. The voice record can be analyzed against a pre-defined list of specific words or phrases to identify instances of profanity or abusive language used by the agent.
1. A method, comprising:
analyzing, via at least one automatic speech recognition component, at least one voice interaction of at least one agent following at least one script, by:
dividing the at least one voice interaction into viewable panel-level segments;
assigning a score to each of the panel-level segments, each score indicating a match accuracy between the panel-level segment and a corresponding expected text of the panel-level segment; and
evaluating each score against a standard, the standard defining a required score for each of the panel-level segments to be declared as a match to their corresponding expected texts.
2. The method of claim 1 , wherein the at least one script includes text corresponding to at least one offer of at least one of goods and services.
3. The method of claim 1 , wherein the at least one voice interaction is conducted via at least one of:
a communications network;
a publicly switched telephone network;
Internet; and
a communications network having at least one wireless component.
4. The method of claim 1 , comprising:
converting the at least one voice interaction into at least one digital signal comprising at least one spectral representation of the at least one voice interaction,
comparing the at least one digital signal to at least one reference standard that includes at least one known vocabulary, and
matching the at least one digital signal to at least one of words and phrases contained in the at least one reference standard.
5. The method of claim 1 , comprising performing at least one action based upon at least one result of the analyzing, wherein the performing includes at least one of:
transmitting at least one signal to the at least one agent;
transmitting at least one signal to at least one reviewing authority; and
making at least one entry in at least one script compliance incentive system.
6. The method of claim 1 , comprising comparing data representing an actual duration of at least one interaction to data representing an expected duration parameter associated with the at least one interaction.
7. The method of claim 1 , comprising determining whether the at least one agent has adequately followed the at least one script by applying a set of action rules to an output of the analyzing, wherein the set of action rules includes a quality assurance action to be taken.
8. The method of claim 1 , comprising comparing each panel-level segment with a corresponding expected text, wherein a confidence level threshold of the automatic speech recognition component is used to evaluate an accuracy of each panel-level segment based on an output of a comparison between each panel-level segment and its corresponding expected text.
9. A non-transitory computer readable medium comprising instructions for:
analyzing, via at least one automatic speech recognition component, at least one voice interaction of at least one agent following at least one script, by:
dividing the at least one voice interaction into viewable panel-level segments;
assigning a score to each of the panel-level segments, each score indicating a match accuracy between the panel-level segment and a corresponding expected text of the panel-level segment; and
evaluating each score against a standard, the standard defining a required score for each of the panel-level segments to be declared as a match to their corresponding expected texts.
10. The non-transitory computer readable medium of claim 9 , comprising:
converting the at least one voice interaction into at least one digital signal comprising at least one spectral representation of the at least one voice interaction,
comparing the at least one digital signal to at least one reference standard that includes at least one known vocabulary, and
matching the at least one digital signal to at least one of words and phrases contained in the at least one reference standard.
11. The non-transitory computer readable medium of claim 9 , comprising performing at least one action based upon at least one result of the analyzing, wherein the performing includes at least one of:
transmitting at least one signal to the at least one agent;
transmitting at least one signal to at least one reviewing authority; and
making at least one entry in at least one script compliance incentive system.
12. The non-transitory computer readable medium of claim 9 , comprising comparing data representing an actual duration of at least one interaction to data representing an expected duration parameter associated with the at least one interaction.
13. The non-transitory computer readable medium of claim 9 , comprising determining whether the at least one agent has adequately followed the at least one script by applying a set of action rules to an output of the analyzing, wherein the set of action rules includes a quality assurance action to be taken.
14. The non-transitory computer readable medium of claim 9 , comprising comparing each panel-level segment with a corresponding expected text, wherein a confidence level threshold of the automatic speech recognition component is used to evaluate an accuracy of each panel-level segment based on an output of a comparison between each panel-level segment and its corresponding expected text.