User: Jnabuurs

The following is an example of a basic business scenario when QnA Service is fully integrated with Agent Assist. QnA ResponseswConfidence.png Assumptions

   Documents have been uploaded to the knowledge base and the models have been successfully trained.
   QnA Service is integrated with Agent Assist.
   Agent Assist is enabled.
   A bot has been configured and deployed.

Business flow

   A bot monitors conversations between a customer and a human agent.
   The customer asks the agent "Can you tell me more about your insurance policies?"
   The bot passes the question to QnA Service.
   QnA Service returns three relevant answers to the bot, each containing a different confidence score.
   The bot displays these answers and confidence scores in the Agent Desktop Application using Agent Assist
   The agent scans the answers and selects the result that best answers the customer's question.
   The agent uses the feedback mechanism (thumbs up/thumbs down) to indicate the accuracy of all three responses. This feedback is sent back to QnA Service, and the confidence scores are updated for subsequent searches.