Proactive knowledge offering system and method
View Patent ↗A system and method for proactively making knowledge offers. A processor is configured to gather information on interactions by a user with resources provided by an enterprise having a customer contact center. The processor anticipates need of the user based on the gathered information, and generates a query based on the anticipated need. Prior to the user expressly requesting knowledge relating to a particular topic, the processor proactively identifies and suggests the knowledge to the user based on the generated query. The processor receives feedback relating to the suggested knowledge and outputs based on the feedback, a relevance score for the suggested knowledge.
1. A method for proactively making a knowledge offer comprising:
gathering, by a processor, information on interactions by a user with resources provided by an enterprise having a customer contact center;
anticipating, by the processor, need of the user based on the gathered information;
generating a search query based on the anticipated need, wherein the generating of the query includes receiving a first query and transforming the first query to a second query;
prior to the user expressly requesting knowledge relating to a particular topic, proactively identifying and suggesting the knowledge to the user, by the processor, based on the generated search query, wherein the proactively identifying and suggesting the knowledge further includes:
retrieving a classifier model for the knowledge;
determining a relevance score based on the classifier model, wherein the relevance score is indicative of a probability that the search query is relevant to the knowledge; and
outputting the suggested knowledge based on the relevance score;
receiving, by the processor, feedback relating to the suggested knowledge;
updating, based on the feedback, the relevance score for the suggested knowledge; and
adding, by the processor, the search query as one of a positive sample or a negative sample for the classifier model, based on the feedback.
2. The method of claim 1 , wherein the resources include a website provided by the enterprise.
3. The method of claim 2 , wherein the interaction is a query typed by the user.
4. The method of claim 1 , wherein the resources include agents of the contact center.
5. The method of claim 1 wherein the interactions include past interactions.
6. The method of claim 1 , wherein the anticipating the need of the user includes anticipating, by the processor, an express inquiry from the user.
7. The method of claim 6 , wherein the information includes context of the interactions.
8. The method of claim 7
wherein the classifier model includes query samples for which the particular document was tagged as being relevant or not relevant.
9. The method of claim 1 , wherein the anticipating the need of the user includes anticipating user intent for a current interaction with one of the resources provided by the enterprise.
10. The method of claim 1 wherein the search query includes text retrieved from the gathered information.
11. The method of claim 1 , wherein the generating of the search query includes:
identifying, by the processor, terms for describing the anticipated need.
12. The method of claim 1 further comprising:
conducting, by the processor, a first level search for a candidate set of knowledge documents based on the knowledge set, and outputting a first score for each of the documents in the candidate set;
conducting, by the processor, a second level search of the candidate set of knowledge documents for relevant documents, and outputting a second score for each of the documents in the candidate set, wherein the second score is the relevance score;
generating, by the processor, a third score as a function of the first and second scores: and
outputting, by the processor, one or more of the documents in the candidate set based on the third score.
13. The method of claim 12 , wherein the feedback is for the particular document, the method further comprising:
training, by the processor, the classifier model for the particular document based on the feedback.
14. The method of claim 1 , wherein the feedback is an express indication by the use as to whether the suggested knowledge was helpful or not.
15. The method of claim 1 , wherein the feedback is implied from actions by the user.
16. The method of claim 1 , wherein the first query includes text input by a user, and the transforming of the query includes expanding the first query with additional text.
17. A system for proactively making a knowledge offer comprising:
processor; and
memory, wherein the memory includes instructions that, when executed by the processor, cause the processor to:
gather information on interactions by a user with resources provided by an enterprise having a customer contact center;
anticipate need of the user based on the gathered information;
generate a search query based on the anticipated need, wherein the generating of the query includes receiving a first query and transforming the first query to a second query;
prior to the user expressly requesting knowledge relating to a particular topic, proactively identify and suggest the knowledge to the user based on the generated search query, wherein the instructions that cause the processor to proactively identify and suggest the knowledge further includes instructions that cause the processor to:
retrieve a classifier model for the knowledge;
determine a relevance score based on the classifier model, wherein the relevance score is indicative of a probability that the search query is relevant to the knowledge; and
output the suggested knowledge based on the relevance score;
receive feedback relating to the suggested knowledge; and
update, based on the feedback, the relevance score for the suggested knowledge; and
adding, by the processor, the search query as one of a positive sample or a negative sample for the classifier model, based on the feedback.