System and method for assisting agents via artificial intelligence
A system and method for handling interactions of a contact center. A processor of the system monitors a real-time interaction between a contact center agent and a user. The monitoring may include analyzing a first input by the user, and a first response by the contact center agent to the first input. The processor identifies an intent of the user based on monitoring the real-time interaction, and in response to identifying the intent, monitors for a first trigger condition. In response to identifying the first trigger condition, the processor invokes an automated agent for outputting a second response to a second input provided by the user. The first trigger condition may be a command to invoke the automated agent, or deviance by the agent from dialog in a dialog script associated with the identified intent.
1. A system for handling interactions comprising:
a processor; and
a memory, wherein the memory stores instructions that, when executed by the processor, cause the processor to:
monitor a real-time interaction between a contact center agent and a user, wherein the monitoring includes analyzing a first input by the user and a first response by the contact center agent to the first input;
identify an intent of the user based on monitoring the real-time interaction;
in response to identifying the intent, monitor for a first trigger condition;
in response to identifying the first trigger condition, invoke an automated agent for outputting a second response to a second input provided by the user, wherein the first trigger condition is a receipt of a command by the contact center agent to invoke the automated agent to engage in an automated conversation with the user instead of the agent, wherein the second response is provided to the user by the automated agent;
monitor for a second trigger condition;
in response to identifying the second trigger condition, disable the automated agent from engaging in the automated conversation with the user for enabling a third response by the contact center agent instead of the automated agent; and
resume monitoring for the first trigger condition.
2. The system of claim 1 , wherein the real-time interaction is at least one of a text-based chat interaction or voice interaction.
3. The system of claim 1 , wherein the instructions further cause the processor to dynamically display on a device of the contact center agent, a user-selectable icon for invoking automated agent.
4. The system of claim 1 , wherein the instructions further cause the processor to:
identify a dialog script, wherein the first trigger condition is identification of deviance by the contact center agent from dialog in the dialog script, wherein the automated agent is configured to provide the second response to the contact center agent as a suggested response.
5. The system of claim 1 , wherein the instructions further cause the processor to:
determine a level of confidence for the determined intent, wherein the first trigger is identification that the level of confidence is above a preset level.
6. The system of claim 1 , wherein the instructions further cause the processor to identify key words provided by the user during the real-time interaction, and provide visual feedback for the identified key words.
7. The system of claim 1 , wherein the instructions that cause the processor to identify the intent include instructions that cause the processor to:
identify key words in the first input and in the first response;
identify a source of each of the key words as being one of the user or the contact center agent; and
perform intent classification based on the identified key words and the identified sources.
8. The system of claim 7 , wherein the instructions that cause the processor to perform the intent classification include instructions that cause the processor to assign a probability value to an identified intent, wherein the probability value depends on a combination of the identified key words and the identified sources.
9. A method for handling interactions comprising:
monitoring, by a processor, a real-time interaction between a contact center agent and a user, wherein the monitoring includes analyzing a first input by the user and a first response by the contact center agent to the first input;
identifying, by the processor, an intent of the user based on monitoring the real-time interaction;
in response to identifying the intent, monitoring, by the processor, for a first trigger condition;
in response to identifying the first trigger condition, invoking, by the processor, an automated agent for outputting a second response to a second input provided by the user, wherein the first trigger condition is receipt of a command by the contact center agent to invoke the automated agent to engage in an automated conversation with the user instead of the agent, wherein the second response is provided to the user by the automated agent;
monitoring, by the processor, for a second trigger condition;
in response to identifying the second trigger condition, disabling, by the processor, the automated agent from engaging in the automated conversation with the user for enabling a third response by the contact center agent instead of the automated agent; and
resuming, by the processor, monitoring for the first trigger condition.
10. The method of claim 9 , wherein the real-time interaction is at least one of a text-based chat interaction or voice interaction.
11. The method of claim 9 further comprising dynamically displaying on a device of the contact center agent, a user-selectable icon for invoking automated agent.
12. The method of claim 9 further comprising:
identifying, by the processor, a dialog script, wherein the first trigger condition is identification of deviance by the contact center agent from dialog in the dialog script, wherein the automated agent is configured to provide the second response to the contact center agent as a suggested response.
13. The method of claim 9 further comprising:
determining, by the processor, a level of confidence for the determined intent, wherein the first trigger is identification that the level of confidence is above a preset level.
14. The method of claim 9 further comprising:
identifying, by the processor, key words provided by the user during the real-time interaction, and providing visual feedback for the identified key words.
15. The method of claim 9 , further comprising:
identifying, by the processor, key words in the first input and in the first response;
identifying, by the processor, a source of each of the key words as being one of the user or the contact center agent; and
performing, by the processor, intent classification based on the identified key words and the identified sources.
16. The method of claim 15 , wherein the performing of the intent classification assigns a probability value to an identified intent, wherein the probability value depends on a combination of the identified key words and the identified sources.