Intent analysis for call center response generation
AssigneeLiveperson, Inc.
InventorsGlen Clodore, Matthew Smith
Life of this patent
Estimating the term…
Abstract
A system obtains conversation data corresponding to conversations between users and agents of a client. The system identifies a set of intents from the conversations and identifies a set of contexts, explicit elements, and implied elements of these intents. The system identifies actions that can be performed to recognize new explicit and implied elements from new conversations and to address intents in these new conversations. Based on these actions, the system generates a set of recommendations that can be provided to the client. As agents communicate with users, the system monitors adherence to the set of recommendations.
Claims21
A computer-implemented method comprising:
- obtaining an incoming message associated with a customer, wherein the incoming message is obtained during a communications session between the customer and an agent;
- determining an intent associated with the incoming message, wherein the intent is determined based on implicit and explicit elements associated with the incoming message;
- receiving a proposed response to the incoming message, wherein the proposed response is associated with the agent, and wherein the proposed response is obtained in real-time during the communications session;
- training a machine learning algorithm, wherein the machine learning algorithm is trained using elements associated with a set of conversations between customers and agents and corresponding responses, and wherein the machine learning algorithm is trained to evaluate responses to incoming messages to improve a likelihood of a positive reception to the responses;
- evaluating the proposed response using the machine learning algorithm to generate feedback for improving the proposed response according to the intent associated with the incoming message;
- and
- dynamically monitoring adherence to the feedback in real-time as messages are exchanged over the communications session.
A system, comprising:
- one or more processors;
- and
- memory storing thereon instructions that, as a result of being executed by the one or more processors, cause the system to:
- obtain an incoming message associated with a customer, wherein the incoming message is obtained during a communications session between the customer and an agent;
- determine an intent associated with the incoming message, wherein the intent is determined based on implicit and explicit elements associated with the incoming message;
- receive a proposed response to the incoming message, wherein the proposed response is associated with the agent, and wherein the proposed response is obtained in real-time during the communications session;
- train a machine learning algorithm, wherein the machine learning algorithm is trained using elements associated with a set of conversations between customers and agents and corresponding responses, and wherein the machine learning algorithm is trained to evaluate responses to incoming messages to improve a likelihood of a positive reception to the responses;
- evaluate the proposed response using the machine learning algorithm to generate feedback for improving the proposed response according to the intent associated with the incoming message;
- and
- dynamically monitor adherence to the feedback in real-time as messages are exchanged over the communications session.
A non-transitory, computer-readable storage medium, storing thereon executable instructions that, as a result of being executed by one or more processors of a computer system, cause the computer system to:
- obtain an incoming message associated with a customer, wherein the incoming message is obtained during a communications session between the customer and an agent;
- determine an intent associated with the incoming message, wherein the intent is determined based on implicit and explicit elements associated with the incoming message;
- receive a proposed response to the incoming message, wherein the proposed response is associated with the agent, and wherein the proposed response is obtained in real-time during the communications session;
- train a machine learning algorithm, wherein the machine learning algorithm is trained using elements associated with a set of conversations between customers and agents and corresponding responses, and wherein the machine learning algorithm is trained to evaluate responses to incoming messages to improve a likelihood of a positive reception to the responses;
- evaluate the proposed response using the machine learning algorithm to generate feedback for improving the proposed response according to the intent associated with the incoming message;
- and
- dynamically monitor adherence to the feedback in real-time as messages are exchanged over the communications session.
Dependencies are inferred from the claim text.