Real-time agent assistance using real-time automatic speech recognition and behavioral metrics
A method of assisting an agent in real-time includes receiving a call interaction between a customer and an agent; identifying words spoken in the call interaction; providing the words to a behavioral models module; computing a score for a plurality of behavioral metrics; providing a phrase formed by the words to a knowledge article selection module; providing each score for the plurality of behavioral metrics to the knowledge article selection module; providing a plurality of knowledge selection rules to the knowledge article selection module; evaluating a combination of the phrase and the scores of the plurality of behavioral metrics against each of the plurality of knowledge selection rules; matching a knowledge selection rule to the combination; selecting a knowledge article associated with the matched knowledge selection rule; generating a visual representation of the selected knowledge article; and presenting in real-time the visual representation on a graphical user interface.
1. A real-time agent assistance system comprising:
a processor and a computer readable medium operably coupled thereto, the computer readable medium comprising a plurality of instructions stored in association therewith that are accessible to, and executable by, the processor, to perform operations which comprise:
receiving a call interaction between a customer and an agent;
providing, by a word spotting module to a knowledge selection module, a phrase formed by words spoken in the call interaction;
providing, by a behavioral models module to a knowledge article selection module, a score for each of a plurality of behavioral metrics that are based at least in part on the words spoken in the call interaction;
providing a plurality of knowledge selection rules to the knowledge article selection module, wherein each knowledge selection rule is associated with a knowledge article;
evaluating, by the knowledge article selection module, a combination of the phrase and the scores for the plurality of behavioral metrics against each of the plurality of knowledge selection rules;
matching, by the knowledge article selection module, a knowledge selection rule to the combination based on the evaluation;
selecting, by the knowledge article selection module, a knowledge article associated with the matched knowledge selection rule; and
presenting, in real-time, by a knowledge article presentation module to the agent or an agent supervisor, a representation of the selected knowledge article on a graphical user interface.
2. The real-time agent assistance system of claim 1 , wherein the operations further comprise:
identifying, by the word spotting module, the words spoken in the call interaction;
providing, by the word spotting module to the behavioral models module, the words spoken in the call interaction; and
computing, by the behavioral models module, the score for each of the plurality of behavioral metrics.
3. The real-time agent assistance system of claim 2 , wherein computing the score for each of the plurality of behavioral metrics comprises:
processing a first word to generate an initial score for each of the plurality of behavioral metrics; and
processing one or more additional words to iteratively update the initial scores.
4. The real-time agent assistance system of claim 1 , wherein the operations further comprise generating, by the knowledge article presentation module, a representation that comprises a visual representation.
5. The real-time agent assistance system of claim 1 , wherein the plurality of behavioral metrics comprises two or more of sentiment, active listening, empathy, demonstration of ownership, building rapport, setting expectations, effective questioning, promotion of self-service, speech velocity, or interruption.
6. The real-time agent assistance system of claim 1 , wherein the operations further comprise identifying time locations in the call interaction at which a spoken instance of a word occurred.
7. The real-time agent assistance system of claim 6 , wherein the operations further comprise determining whether the agent or the customer spoke the word.
8. The real-time agent assistance system of claim 1 , wherein the operations further comprise receiving the plurality of knowledge selection rules from a user.
9. The real-time agent assistance system of claim 8 , wherein the operations further comprise:
receiving an updated plurality of knowledge selection rules from the user;
re-evaluating, by the knowledge article selection module, the combination of the phrase and the scores for the plurality of behavioral metrics against each of the updated plurality of knowledge selection rules;
re-matching, by the knowledge article selection module, an updated knowledge selection rule to the combination based on the evaluation; and
re-selecting, by the knowledge article selection module, a knowledge article associated with the updated matched knowledge selection rule.
10. The real-time agent assistance system of claim 4 , wherein the visual representation comprises an alert, a message, a score, or a combination thereof.
11. A method of assisting an agent in real-time, which comprises:
receiving a call interaction between a customer and an agent;
providing, by a word spotting module to the knowledge selection module, a phrase formed by words spoken in the call interaction to a knowledge selection module;
providing, by a behavioral models module to a knowledge article selection module, a score for each of a plurality of behavioral metrics that are based at least in part on the words spoken in the call interaction;
providing a plurality of knowledge selection rules to the knowledge article selection module, wherein each knowledge selection rule is associated with a knowledge article;
evaluating, by the knowledge article selection module, a combination of the phrase and the scores for the plurality of behavioral metrics against each of the plurality of knowledge selection rules;
matching, by the knowledge article selection module, a knowledge selection rule to the combination based on the evaluation;
selecting, by the knowledge article selection module, a knowledge article associated with the matched knowledge selection rule; and
presenting, in real-time, by a knowledge article presentation module to the agent or an agent supervisor, a visual representation of the selected knowledge article on a graphical user interface.
12. The method of claim 11 , wherein the plurality of behavioral metrics comprise two or more of sentiment, active listening, empathy, demonstration of ownership, building rapport, setting expectations, effective questioning, promotion of self-service, speech velocity, or interruption.
13. The method of claim 11 , further comprising identifying time locations in the call interaction at which a spoken instance of a word occurred.
14. The method of claim 13 , further comprising determining whether the agent or the customer spoke the word.
15. The method of claim 11 , further comprising receiving the plurality of knowledge selection rules from a user.
16. A non-transitory computer-readable medium having stored thereon computer-readable instructions executable by a processor to perform operations which comprise:
receiving a call interaction between a customer and an agent;
providing, by a word spotting module to a knowledge selection module, a phrase formed by words spoken in the call interaction;
providing, by a behavioral models module to a knowledge article selection module, a score for each of a plurality of behavioral metrics that are based at least in part on the words spoken in the call interaction;
providing a plurality of knowledge selection rules to the knowledge article selection module, wherein each knowledge selection rule is associated with a knowledge article;
evaluating, by the knowledge article selection module, a combination of the phrase and the scores for the plurality of behavioral metrics against each of the plurality of knowledge selection rules;
matching, by the knowledge article selection module, a knowledge selection rule to the combination based on the evaluation;
selecting, by the knowledge article selection module, a knowledge article associated with the matched knowledge selection rule; and
presenting, in real-time, by a knowledge article presentation module to the agent or an agent supervisor, a visual representation of the selected knowledge article on a graphical user interface.
17. The non-transitory computer-readable medium of claim 16 , wherein the plurality of behavioral metrics comprise two or more of sentiment, active listening, empathy, demonstration of ownership, building rapport, setting expectations, effective questioning, promotion of self-service, speech velocity, or interruption.
18. The non-transitory computer-readable medium of claim 16 , wherein the operations further comprise identifying time locations in the call interaction at which a spoken instance of a word occurred, and determining whether the agent or the customer spoke the word.
19. The non-transitory computer-readable medium of claim 16 , wherein the operations further comprise receiving the plurality of knowledge selection rules from a user.
20. The non-transitory computer-readable medium of claim 19 , wherein the operations further comprise:
receiving an updated plurality of knowledge selection rules from the user;
re-evaluating, by the knowledge article selection module, the combination of the phrase and the scores for the plurality of behavioral metrics against each of the updated plurality of knowledge selection rules;
re-matching, by the knowledge article selection module, an updated knowledge selection rule to the combination based on the evaluation; and
re-selecting, by the knowledge article selection module, a knowledge article associated with the updated matched knowledge selection rule.