Real-time agent assistance
View Patent ↗Some general aspects of the invention relate to systems and methods for improving contact center agent performance, for instance, by integrating real-time call monitoring with speech analytics to present agents with information useful to the handling of the current calls. In some implementations, phonetically based speech analysis techniques are applied to process live audio streams to identify key words and/or phrases of relevance, based on which knowledge articles can be selectively presented to agents to drive more efficient business processes.
1. A method comprising:
accepting a specification of a predefined plurality of associations between keyphrases of a predefined set of keyphrases and knowledge articles of a predefined set of knowledge articles stored in a machine-readable data store;
processing, by a word spotting engine of a contact center system, audio data representative of an interaction between a contact center caller and a contact center agent to form an output including identifications of occurrences of keyphrases of the predefined set of keyphrases in the audio data;
selecting, from the predefined set of knowledge articles, a candidate set of knowledge articles for presentation to the contact center agent based at least in part on a result of processing the identified occurrences of the keyphrases in the output of the word spotting engine according to the predefined plurality of associations; and
detecting that the interaction between the contact center agent and the contact center caller is ongoing and causing visual representations of selected knowledge articles of the candidate set to be presented on a display terminal associated with the contact center agent during the ongoing interaction; wherein each of the selected knowledge articles of the candidate set is associated with a relevance score that exceeds a threshold score, and wherein the method further comprises:
comparing the relevance scores associated with the selected knowledge articles of the candidate set to identify the knowledge article of the candidate set that is most relevant to the interaction between the contact centre caller and the contact center agent; and
generating a visual representation of the identified knowledge article of the candidate set.
2. The method of claim 1 , wherein processing the audio data comprises:
identifying time locations in the audio data at which a spoken instance of a keyphrase of the predefined set of keyphrases is likely to have occurred based on a comparison of data representing the audio data with data representing spoken instances of at least some of the keyphrases of the predefined set.
3. The method of claim 1 , further comprising:
assigning relevance scores to knowledge articles of the predefined set, each relevance score representing a measure of relevance of its respective knowledge article to the occurrences of one or more keyphrases;
designating a knowledge article of the predefined set as belonging to a candidate set of knowledge articles based on a comparison of its respective relevance score to a threshold score.
4. The method of claim 3 , further comprising:
generating a relevance score to be assigned to a knowledge article of the predefined set, wherein the generating includes:
processing a first segment of the audio data to generate an initial relevance score; and
processing one or more additional segments of the audio data to iteratively update the initial relevance score.
5. The method of claim 1 , further comprising:
generating visual representations of selected knowledge articles of the candidate set.
6. The method of claim 1 , further comprising:
prior to comparing the relevance scores associated with the selected knowledge articles of the candidate set,
processing audio data representative of an immediately prior portion of the interaction between the contact center caller and the contact center agent; and
updating the relevance score associated with each of the selected knowledge articles of the candidate set based on a result of the processing.
7. The method of claim 1 , wherein the predefined set of knowledge articles includes content associated with promotions, incentives, products, product accessories, services, service enhancements, and agent-assisting content.
8. The method of claim 1 , further comprising:
receiving user input indicative of a perceived relevance of at least one of the selected knowledge articles that is presented on the display terminal associated with the contact center agent.
9. The method of claim 1 , further comprising:
analyzing a plurality of user inputs each indicative of a perceived relevance of a first of the selected knowledge articles that is presented on the display terminals associated with respective contact center agents; and
based on results of the analyzing, adjusting selection criteria for subsequent use in selecting, from the predefined set of knowledge articles stored in a machine-readable data store, a candidate set of knowledge articles for presentation.
10. The method of claim 1 wherein each association of the predefined plurality of associations includes a quantitative score representing a degree of importance of a keyphrase to a knowledge article, and processing the identified occurrences of the keyphrases in the output of the word spotting engine according to the predefined plurality of associations utilizes the quantitative scores.
11. The method of claim 1 wherein at least some of the steps are implemented in a contact center service platform running on a call center switching system.