Computer-implemented system and method for facilitating interactions via automatic agent responses
A computer-implemented system and method for facilitating interactions via automatic agent responses is provided. Communication during an interaction between a user and an agent is monitored and a request from the user is identified during the interaction. A list of candidate responses to the request is compiled and provided to the agent. Time is measured upon providing the list to the agent and a predetermined amount of time is applied to the measured time. When the measured time exceeds the predetermined amount of time and the agent has failed to provide a response to the user, one of the candidate responses from the list is automatically selected and provided to the user.
1. A computer-implemented system for facilitating interactions via automatic agent responses, comprising:
a communication monitor to monitor communication during an interaction between a user and an agent;
a request identification module to identify a request from the user during the interaction;
a compiler to compile a list of candidate responses to the request;
a delivery module to provide the list of candidate responses to the agent for selecting at least one of the responses;
a time application module to count time upon providing the list to the agent and to apply a predetermined time to the counted time;
a selection module to automatically select one of the candidate responses from the list when the agent fails to perform at least one of selecting one or more of the responses from the list and responding to the user within the predetermined time as determined by the counted time; and
a response delivery module to provide the selected candidate response to the user.
2. A system according to claim 1 , further comprising:
a list generator to select the candidate responses for inclusion in the list via at least one of response models and machine learning.
3. A system according to claim 1 , further comprising:
a response identification module to identify the candidate responses based on one or more of a likelihood that each candidate response satisfies the request and a sentiment of the user regarding each candidate response for the request.
4. A system according to claim 3 , further comprising:
a sentiment determination module to determine the sentiment of the user to one or more of the candidate responses by providing, during a prior interaction, the user with a response from the list of candidate responses and by identifying the sentiment of the user based on the response.
5. A system according to claim 4 , wherein the response is selected from another list of candidate responses for a previous request made by the user or received directly from the agent or a different agent.
6. A system according to claim 4 , further comprising:
a sentiment calculation module to measure the sentiment of the user, comprising:
a whisper module to provide a whisper to the user regarding the sentiment of the user to the response; and
a receipt module to receive from the user a reply comprising the user's sentiment.
7. A system according to claim 6 , wherein the selected candidate response is associated with a high or positive sentiment measure.
8. A system according to claim 3 , further comprising:
a likelihood determination module to determine a likelihood that each candidate response satisfies the request based on one or more similarity factors comprising whether each such candidate response was provided to a related request in one or more previous interactions and a number of times each candidate response was provided to the agent in reply to the related requests; and
a likelihood calculation module to assign a measure of likelihood to each candidate response based on the factors.
9. A system according to claim 8 , wherein the selected candidate response is associated with a higher or highest likelihood measure.
10. A computer-implemented method for facilitating interactions via automatic agent responses, comprising:
monitoring communication during an interaction between a user and an agent;
identifying a request from the user during the interaction;
compiling a list of candidate responses to the request;
providing the list of candidate responses to the agent for selecting at least one of the responses;
counting time upon providing the list to the agent;
applying a predetermined time to the counted time;
automatically selecting one of the candidate responses from the list when the agent fails to perform at least one of selecting one or more of the responses from the list and responding to the user within the predetermined time as determined by the counted time; and
providing the selected candidate response to the user.
11. A method according to claim 10 , further comprising:
selecting the candidate responses via at least one of response models and machine learning.
12. A method according to claim 10 , further comprising:
identifying the candidate responses based on one or more of a likelihood that each candidate response satisfies the request and a sentiment of the user regarding each candidate response for the request.
13. A method according to claim 12 , further comprising:
determining the sentiment of the user to one or more of the candidate responses, comprising:
providing, during a prior interaction, the user with a response from the list of candidate responses; and
identifying the sentiment of the user based on the response.
14. A method according to claim 13 , wherein the response is selected from another list of candidate responses for a previous request made by the user or received directly from the agent or a different agent.
15. A method according to claim 13 , further comprising:
measuring the sentiment of the user, comprising:
providing a whisper to the user regarding the sentiment of the user to the response; and
receiving from the user a reply comprising the user's sentiment.
16. A method according to claim 15 , wherein the selected candidate response is associated with a high or positive sentiment measure.
17. A method according to claim 12 , further comprising:
determining a likelihood that each candidate response satisfies the request based on one or more similarity factors comprising whether each such candidate response was provided to a related request in one or more previous interactions and a number of times each candidate response was provided to the agent in reply to the related requests; and
assigning a measure of likelihood to each candidate response based on the factors.
18. A method according to claim 17 , wherein the selected candidate response is associated with a higher or highest likelihood measure.
19. A method according to claim 10 , wherein the list of candidate responses is customized for the user by revising the list of candidate responses based on a reaction of the user with respect to one or more of the candidate responses previously provided to the user.
20. A system according to claim 1 , wherein the list of candidate responses is customized for the user by revising the list of candidate responses based on a reaction of the user with respect to one or more of the candidate response previously provided to the user.