Method and system for identifying ideal virtual assistant bots for providing response to user queries
A method and bot assistance system ( 101 ) for identifying ideal virtual assistant bots for providing response to user queries is disclosed. The bot assistance system receives user query. The user query is processed to identify context and intent using Natural Language Processing (NLP). The bot assistance system identifies at least one virtual assistant bot for responding to user query based on confidence score of each virtual assistant bot associated with intent of user query, using pretrained model. The identification of at least one virtual assistant bot includes either determining ideal virtual assistant bot based on context of user query, using predefined governance rules. Otherwise, initiating polling between plurality of predefined virtual assistant bots when context of user query is unidentified. The plurality of predefined virtual assistance bots is provided to user associated with user query if each of predefined virtual assistance bots are selected during polling.
1. A method of identifying ideal virtual assistant bots for providing response to user queries, the method comprising:
receiving, by a bot assistance system, a user query from a user device, wherein the user query is processed to identify a context and intent using Natural Language Processing (NLP);
identifying, by the bot assistance system, at least one virtual assistant bot from a plurality of virtual assistant bots associated with an entity for responding to the user query, based on a confidence score of each virtual assistant bot associated with the intent of the user query, using a pretrained model, wherein the identification of at least one virtual assistant bot comprises:
determining an ideal virtual assistant bot based on the context of the user query, using predefined governance rules; or
initiating a polling between a plurality of predefined virtual assistant bots of the plurality of virtual assistant bots, when the context of the user query is unidentified, wherein the plurality of predefined virtual assistance bots is provided to a user associated with the user query, if each of the plurality of predefined virtual assistance bots are selected during polling; and
wherein identifying at least one virtual assistant bot from the plurality of virtual assistant bots comprises comparing the confidence score associated with the intent of each virtual assistant bot with a threshold confidence score, and wherein when the confidence score is below the threshold confidence score, providing to a user, a list of responses closely matching to the user query using an unsupervised fallback model.
2. The method as claimed in claim 1 , further comprising storing information about at least one response from the list of responses selected by the user.
3. The method as claimed in claim 1 , further comprising setting the plurality of virtual assistant bots as primary bots and secondary bots, wherein the primary bots are the predefined virtual assistant bots which are authorised for participating in the polling and the secondary bots are associated with selection by the bot assistance system.
4. The method as claimed in claim 1 , further comprises setting a priority for each of the predefined virtual assistant bots.
5. The method as claimed in claim 1 , further comprising initiating a feedback learning for a fallback model using suggestions provided by a number of users for a similar user query.
6. A bot assistance system for identifying ideal virtual assistant bots for providing response to user queries, comprising:
a processor; and
a memory communicatively coupled to the processor, wherein the memory stores processor instructions, which, on execution, causes the processor to:
receive a user query from a user device, wherein the user query is processed to identify a context and intent using Natural Language Processing (NLP);
identify at least one virtual assistant bot from a plurality of virtual assistant bots associated with an entity for responding to the user query, based on a confidence score of each virtual assistant bot associated with the intent of the user query, using a pretrained model, wherein the identification of at least one virtual assistant bot comprises:
determining an ideal virtual assistant bot based on the context of the user query, using predefined governance rules; or
initiating a polling between a plurality of predefined virtual assistant bots of the plurality of virtual assistant bots, when the context of the user query is unidentified, wherein the plurality of predefined virtual assistance bots is provided to a user associated with the user query, if each of the plurality of predefined virtual assistance bots are selected during polling; and
wherein the processor identifies the at least one virtual assistant bot by comparing the confidence score associated with the intent of each virtual assistant bot with a threshold confidence score, and wherein when the confidence score is below the threshold confidence score, the processor provides a list of responses closely matching to the user query using an unsupervised fallback model.
7. The bot assistance system as claimed in claim 6 , wherein the processor initiates storing information about at least one response from the list of responses selected by the user.
8. The bot assistance system as claimed in claim 6 , wherein the processor sets the plurality of virtual assistant bots as primary bots and secondary bots, wherein the primary bots are the predefined virtual assistant bots which are authorised for participating in the polling and the secondary bots are associated with selection by the bot assistance system.
9. The bot assistance system as claimed in claim 6 , wherein the processor sets a priority for each of the predefined virtual assistant bots.
10. The bot assistance system as claimed in claim 6 , wherein the processor initiates a feedback learning for a fallback model using suggestions provided by a number of users for a similar user query.