Systems and methods for training a virtual assistant platform configured to assist human agents at a contact center
A contact center server provides to an agent device one or more automated response recommendations determined by an executable virtual assistant platform to correspond to customer message data received from a customer device as part of conversation data. Further, the contact center server receives a selection of one of the automated response recommendations from the agent device and identifies agent response data to the customer message data transmitted from the agent device to the customer device. Further, the contact center server, using one or more classification models, determines when there is a change between the selected response recommendation and the identified agent response data. Further, the contact center server associates one or more tags to the identified agent response data in the conversation data when the determination indicates the change. Subsequently, the contact center server updates training of the executable virtual assistant platform based on the conversation data with the associated one or more tags.
1. A method comprising:
providing to an agent device, by a contact center server, recommendation data comprising one or more automated response recommendations determined by an executable virtual assistant platform to correspond to customer message data received from a customer device as part of conversation data;
determining, by the contact center server, using one or more classification models, when there are one or more changes between the one or more automated response recommendations and an agent response data to the customer message data sent as part of the conversation data;
associating, by the contact center server, one or more tags to the agent response data in the conversation data when the determination indicates the one or more changes; and
updating training, by the contact center server, of the executable virtual assistant platform based on the conversation data with the associated one or more tags.
2. The method of claim 1 , wherein the conversation data between the customer device and the agent device comprises text-based data or voice-based data.
3. The method of claim 1 , wherein the one or more automated response recommendations in the recommendation data comprise at least one of: one or more responses; or one or more snippets from one or more knowledge articles.
4. The method of claim 1 , wherein the determining when there are the one or more changes further comprises identifying at least one of: rephrasing the one or more automated response recommendations; adding new content to the one or more automated response recommendations; modifying content of the one or more automated response recommendations; or deleting content from the one or more automated response recommendations.
5. The method of claim 4 , wherein the identifying the rephrasing the one or more automated response recommendations in the recommendation data further comprises identifying at least one of changing tone or grammar of the one or more automated response recommendations in the recommendation data.
6. The method of claim 4 , wherein the identifying the adding of new content to the one or more automated response recommendations in the recommendation data further comprises identifying at least one of: adding one or more intents; adding one or more entities; adding one or more entity values; adding empathy; adding small-talk; or adding a greeting.
7. The method of claim 4 , wherein the identifying the modifying of content of the one or more automated response recommendations in the recommendation data further comprises at least one of: modifying one or more intents; modifying one or more entities; or modifying one or more entity values.
8. The method of claim 4 , wherein the identifying the deleting content from the one or more automated response recommendations in the recommendation data further comprises at least one of: deleting one or more intents; deleting one or more entities; deleting one or more entity values; deleting empathy; deleting small-talk; deleting a greeting; or deleting a portion of the one or more automated response recommendations.
9. The method of claim 1 , wherein the one or more classification models comprise: an empathy detection model; an intent and entity model; or a small-talk and greeting model.
10. The method of claim 1 , further comprising:
prior to the updating the training of the executable virtual assistant platform:
sending, by the contact center server, a notification to an enterprise user device about the determined one or more changes; and
receiving, by the contact center server, an approval from the enterprise user device to update the training of the executable virtual assistant platform based on the conversation data with the associated one or more tags.
11. A contact center server comprising:
one or more processors; and
a memory coupled to the one or more processors which are configured to execute programmed instructions stored in the memory to:
provide to an agent device recommendation data comprising one or more automated response recommendations determined by an executable virtual assistant platform to correspond to customer message data received from a customer device as part of conversation data;
determine using one or more classification models, when there are one or more changes between the one or more automated response recommendations and an agent response data to the customer message data sent as part of the conversation data;
associate one or more tags to the agent response data in the conversation data when the determination indicates the one or more changes; and
update training of the executable virtual assistant platform based on the conversation data with the associated one or more tags.
12. The contact center server of claim 11 , wherein the conversation data between the customer device and the agent device comprises text-based data or voice-based data.
13. The contact center server of claim 11 , wherein the one or more automated response recommendations in the recommendation data comprise at least one of: one or more responses; or one or more snippets from one or more knowledge articles.
14. The contact center server of claim 11 , wherein to determine the one or more changes when there are the one or more changes, the one or more processors are further configured to identify at least one of: rephrasing the one or more automated response recommendations; adding new content to the one or more automated response recommendations; modifying content of the one or more automated response recommendations; or deleting content from the one or more automated response recommendations.
15. The contact center server of claim 14 , wherein the identifying the rephrasing the one or more automated response recommendations in the recommendation data further comprises identifying at least one of changing tone or grammar of the one or more automated response recommendations in the recommendation data.
16. The contact center server of claim 14 , wherein the identifying the adding of new content to the one or more automated response recommendations in the recommendation data further comprises identifying at least one of: adding one or more intents; adding one or more entities; adding one or more entity values; adding empathy; adding small-talk; or adding a greeting.
17. The contact center server of claim 14 , wherein the identifying the modifying of content of the one or more automated response recommendations in the recommendation data further comprises at least one of: modifying one or more intents; modifying one or more entities; or modifying one or more entity values.
18. The contact center server of claim 14 , wherein the identifying the deleting content from the one or more automated response recommendations in the recommendation data further comprises at least one of: deleting one or more intents; deleting one or more entities; deleting one or more entity values; deleting empathy; deleting small-talk; deleting a greeting; or deleting a portion of the one or more automated response recommendations.
19. The contact center server of claim 11 , wherein the one or more classification models comprise: an empathy detection model; an intent and entity model; or a small-talk and greeting model.
20. The contact center server of claim 11 , wherein the one or more processors, prior to the updating the training of the executable virtual assistant platform, are further configured to execute programmed instructions stored in the memory to:
send a notification to an enterprise user device about the determined one or more changes; and
receive an approval from the enterprise user device to update the training of the executable virtual assistant platform based on the conversation data with the associated one or more tags.
21. A non-transitory computer-readable medium storing instructions which when executed by one or more processors, causes the one or more processors to:
provide to an agent device recommendation data comprising one or more automated response recommendations determined by an executable virtual assistant platform to correspond to customer message data received from a customer device as part of conversation data;
determine using one or more classification models, when there are one or more changes between the one or more automated response recommendations and an agent response data to the customer message data sent as part of the conversation data;
associate one or more tags to the agent response data in the conversation data when the determination indicates the one or more changes; and
update training of the executable virtual assistant platform based on the conversation data with the associated one or more tags.
22. The non-transitory computer-readable medium of claim 21 , wherein the conversation data between the customer device and the agent device comprises text-based data or voice-based data.
23. The non-transitory computer-readable medium of claim 21 , wherein the one or more automated response recommendations in the recommendation data comprise at least one of: one or more responses; or one or more snippets from one or more knowledge articles.
24. The non-transitory computer-readable medium of claim 21 , wherein the determining the one or more changes when there are the one or more changes further comprises identifying at least one of: rephrasing the one or more automated response recommendations; adding new content to the one or more automated response recommendations; modifying content of the one or more automated response recommendations; or deleting content from the one or more automated response recommendations.
25. The non-transitory computer-readable medium of claim 24 , wherein the identifying the rephrasing the one or more automated response recommendations in the recommendation data further comprises identifying at least one of changing tone or grammar of the one or more automated response recommendations in the recommendation data.
26. The non-transitory computer-readable medium of claim 24 , wherein the identifying the adding of new content to the one or more automated response recommendations in the recommendation data further comprises identifying at least one of: adding one or more intents; adding one or more entities; adding one or more entity values; adding empathy; adding small-talk; or adding a greeting.
27. The non-transitory computer-readable medium of claim 24 , wherein the identifying the modifying of content of the one or more automated response recommendations in the recommendation data further comprises at least one of: modifying one or more intents; modifying one or more entities; or modifying one or more entity values.
28. The non-transitory computer-readable medium of claim 24 , wherein the identifying the deleting content from the one or more automated response recommendations in the recommendation data further comprises at least one of: deleting one or more intents; deleting one or more entities; deleting one or more entity values; deleting empathy; deleting small-talk; deleting a greeting; or deleting a portion of the one or more automated response recommendations.
29. The non-transitory computer-readable medium of claim 21 , wherein the one or more classification models comprise: an empathy detection model; an intent and entity model; or a small-talk and greeting model.
30. The non-transitory computer-readable medium of claim 21 , further comprising instructions which when executed by the one or more processors prior to the updating the training of the executable virtual assistant platform, causes the one or more processors to:
send a notification to an enterprise user device about the determined one or more changes; and
receive an approval from the enterprise user device to update the training of the executable virtual assistant platform based on the conversation data with the associated one or more tags.