System and method of sentiment modeling and application to determine optimized agent action
The present invention is a system and method of continuous sentiment tracking and the determination of optimized agent actions through the training of sentiment models and applying the sentiment models to new incoming interactions. The system receives conversations comprising incoming interactions and agent actions and determines customer sentiment on a micro-interaction level for each incoming interaction. Based on interaction types, the system correlates the determined sentiment with the agent action received prior to the sentiment determination to create and train sentiment models. Sentiment models include agent action recommendations for a desired sentiment outcome. Once trained, the sentiment models can be applied to new incoming interactions to provide CSRs with actions that will yield a desired sentiment outcome.
1. A method for automatedly providing optimized agent actions, comprising:
receiving a set of incoming interactions along with a sentiment for each incoming interaction;
receiving an agent action associated with each incoming interaction in the set of incoming interactions, wherein the associated agent action is the agent action that immediately precedes the incoming interaction;
assigning an interaction type to each of the agent actions;
for each interaction type, performing a model analysis of each agent action and a sentiment associated with the agent action;
creating at least one sentiment model for each interaction type based on the model analysis, wherein the sentiment model includes at least one optimized agent action and an interaction type;
applying the sentiment models to a new incoming interaction to determine an optimized agent action for the new incoming interaction; and
displaying to a customer service representative the determined optimized agent action.
2. The method of claim 1 , wherein each interaction of the set of incoming interactions is associated with a conversation.
3. The method of claim 2 , wherein the sentiment associated with each agent action is the sentiment assigned to the incoming interaction immediately following the agent action in the associated conversation.
4. The method of claim 1 , wherein for each interaction type the at least one sentiment model is created if the interaction type does not already have a created sentiment model.
5. The method of claim 1 , wherein for each interaction type, when the interaction type already has a created sentiment model, the sentiment model for the interaction type is updated based on the model analysis for the interaction type.
6. The method of claim 1 , the method further comprising determining the interaction type of the new incoming interaction prior to applying the sentiment models, further wherein the determination of the optimized agent action for the new incoming interaction is based on the determined interaction type of the new incoming interaction.
7. The method of claim 1 , the method further comprising determining the sentiment of the new incoming interaction.
8. A method for processing interactions to provide optimized agent actions, comprising:
providing a customer engagement center (CEC) for processing interactions to create sentiment models and provide at least one optimized agent action in response to a new incoming interaction;
receiving an agent action in response to an interaction;
generating interaction metadata for the agent action based on an analysis of the agent action, wherein the interaction metadata includes an interaction type for the agent action;
receiving an incoming interaction in response to the agent action;
generating a sentiment for the incoming interaction based on an analysis of the incoming interaction, a set of interaction metadata, and a set of sentiment criteria;
performing a model analysis of the agent action along with the corresponding sentiment of the incoming interaction to create at least one sentiment model for the interaction type associated with the agent action based on the model analysis if the interaction type does not already have a sentiment model, wherein the sentiment model includes at least one optimized agent action and one of the interaction types;
updating the sentiment models for the interaction type associated with the agent action based on the model analysis, if the interaction type already has a sentiment model;
applying the at least one sentiment model to the new incoming interaction to determine an optimized agent action; and
displaying to a customer service representative the determined optimized agent action.
9. The method of claim 8 , wherein the interaction, the agent action and the incoming interaction in response to the agent action are all part of a conversation.
10. The method of claim 9 , further comprising continuing to receive additional interactions and additional agent actions, wherein the additional interactions and additional agent actions are part of the conversation, until the conversation is ended.
11. The system of claim 10 , further comprising generating a sentiment for each additional interaction based on the interaction and the set of sentiment criteria.
12. The system of claim 11 , wherein the model analysis is performed for each agent action from the conversation.
13. The system of claim 8 , the method further comprising determining the interaction type for the new incoming interaction.
14. The system of claim 13 , wherein the at least one sentiment model applied to the new incoming interaction is based on the determined interaction type for the new incoming interaction.
15. A method for processing interactions to provide optimized agent actions, comprising:
providing a customer engagement center (CEC) for processing conversations to create sentiment models and provide at least one optimized agent in response to a new incoming interaction;
receiving an agent action at the CEC in response to an interaction;
performing an agent analysis of the agent action to assign an interaction type to the agent action;
receiving an incoming interaction in response to the agent action;
generating a sentiment for the incoming interaction based on an analysis of the incoming interaction, a set of interaction metadata, and a set of sentiment criteria;
determining a coupled agent action, wherein the coupled agent action includes the agent action and the sentiment assigned to the responsive incoming interaction;
performing a model analysis of the coupled agent action for the conversation to create at least one sentiment model for the interaction type associated with the coupled agent action based on the model analysis when the interaction type does not already have a sentiment model, wherein the sentiment model includes at least one optimized agent action and one of the interaction types;
updating the sentiment models for the interaction type associated with the coupled agent action based on the model analysis, when the interaction type already has a sentiment model;
applying the at least one sentiment model to the new incoming interaction to determine an optimized agent action; and
displaying to a customer service representative the determined optimized agent action.
16. The method of claim 15 , wherein the agent action and the incoming interaction are associated with a conversation.
17. The method of claim 16 , further comprising continuing to receive additional agent interactions and additional incoming interactions in response to the additional agent actions, wherein the additional incoming interactions and additional agent actions are part of the conversation, until the conversation is ended.
18. The method of claim 17 , further comprising generating sentiment for each additional incoming interaction based on the additional interaction and the set of sentiment criteria.
19. The method of claim 18 , wherein the coupled agent action is a plurality of coupled agent actions for the conversation, wherein each coupled agent actions includes one of the additional agent actions from the conversation and the sentiment assigned to the additional responsive incoming interaction.
20. The method of claim 15 , the method further comprising determining the sentiment of the new incoming interaction and displaying to the customer service representative the sentiment for the new incoming interaction.