Method and system for optimizing call distribution among callers for improved positive outcome

Patent metadata

Field	Value
Publication number	US-10609216-B1
Application number	US-201916542380-A
Country	US
Kind code	B1
Filing date	Aug 16, 2019
Priority date	Jun 27, 2019
Publication date	Mar 31, 2020
Grant date	Mar 31, 2020

Abstract

A method and a system of optimizing call distribution among callers for improved positive outcome is disclosed. In an embodiment, the method may include receiving at least one attribute associated with one or more callers, and at least one attribute associated with one or more callees. The method may further include calculating an overall cost associated with each of the one or more callers with respect to each of the one or more callees, based on the at least one attribute associated with the one or more callers and the at least one attribute associated with the one or more callees, using an Artificial Intelligence (AI) model. The method may further include allocating a caller to a callee based on the overall cost.

First claim

What is claimed is:

1. A method of optimizing call distribution among callers for improved positive outcome, the method comprising:
receiving, by a call distribution optimization device, at least one attribute associated with one or more callers, and at least one attribute associated with one or more callees;
calculating, by the call distribution optimization device, an overall cost associated with each of the one or more callers with respect to each of the one or more callees, based on the at least one attribute associated with the one or more callers and the at least one attribute associated with the one or more callees, using an Artificial Intelligence (AI) model; and
allocating, by the call distribution optimization device, a caller to a callee based on the overall cost associated with each of the one or more callers with respect to each of the one or more callees.

2. The method of claim 1 , wherein the at least one attribute associated with the one or more callers and the at least one attribute associated with one or more callees comprise callee information, caller information, historical data of previous conversations between callers and callees, callee behavior patterns, caller behavior patterns, and dialogues of the callers in previous conversations.

3. The method of claim 1 , wherein calculating the overall cost comprises:
determining at least one of a propensity score of each of the one or more callees, an effective skillset of each of the one or more callers, and an effective experience of each of the one or more callers, based on the at least one attribute associated with the one or more callers and the at least one attribute associated with the one or more callees; and
calculating the overall cost associated with each of the one or more callers, based on the at least one of the propensity score of each of the one or more callees, the effective skillset of each of the one or more callers, and the effective experience of each of the one or more caller, using the AI model.

4. The method of claim 1 further comprising:
dynamically determining an intent of conversation between the callee and the caller allocated to the callee, using an Artificial Neural Network (ANN) and a Natural Language Processing (NLP) model; and
dynamically generating a conversation template, based on the intent of the conversation between the caller and the callee allocated to the caller, for assisting the caller in achieving an objective of the conversation, using Natural Language Generation (NLG) model.

5. The method of claim 4 , wherein dynamically generating the conversation template further comprises:
dynamically determining temporal goal-specific intents of the conversation between the callee and the caller allocated to the callee in each of one or more timeframes associated with the conversation;
calculating a probability of achieving the objective of the call based on the temporal goal-specific intents of the conversation; and
dynamically generating the conversation template, based on the temporal goal-specific intents of the conversation and the probability of achieving the objective of the call.

6. The method of claim 4 , wherein the ANN is a Long Short-Term Memory Network (LSTM).

7. The method of claim 6 , wherein the ANN is trained with the temporal goal-specific intents of the conversation dynamically determined between each of the one or more callers and the set of callees allocated to each of the one or more callers.

8. The method of claim 1 , wherein the AI model is an Asynchronous Actor Critic Network model.

9. A call distribution optimization device for optimizing call distribution among callers for improved positive outcome, the call distribution optimization device comprising:
a processor; and
a memory communicatively coupled to the processor, wherein the memory stores processor-executable instructions, which, on execution, cause the processor to:
receive at least one attribute associated with one or more callers, and at least one attribute associated with one or more callees;
calculate an overall cost associated with each of the one or more callers with respect to each of the one or more callees, based on the at least one attribute associated with the one or ore callers and the at least one attribute associated with the one or more callees, using an Artificial Intelligence (AI) model; and
allocate a caller to a callee based on the overall cost associated with each of the one or more callers with respect to each of the one or more callees.

10. The call distribution optimization device of claim 9 , wherein the at least one attribute associated with the one or more callers and the at least one attribute associated with one or more callees comprise callee information, caller information, historical data of previous conversations between callers and callees, callee behavior patterns, caller behavior patterns, and dialogues of the callers in previous conversations.

11. The call distribution optimization device of claim 9 , wherein calculating the overall cost comprises:
determining at least one of a propensity score of each of the one or more callees, an effective skillset of each of the one or more callers, and an effective experience of each of the one or more callers, based on the at least one attribute associated with the one or more callers and the at least one attribute associated with the one or more callees; and
calculating the overall cost associated with each of the one or more callers, based on the at least one of the propensity score of each of the one or more callees, the effective skiliset of each of the one or more callers, and the effective experience of each of the one or more caller, using the AI model.

12. The call distribution optimization device of claim 9 , wherein the processor-executable instructions, on execution, further cause the processor to:
dynamically determine an intent of conversation between the callee and the caller allocated to the callee, using an Artificial Neural Network (ANN) and a Natural Language Processing (NLP) model; and
dynamically generate a conversation template, based on the intent of the conversation between the caller and the callee allocated to the caller, for assisting the caller in achieving an objective of the conversation, using Natural Language Generation (NLG) model.

13. The call distribution optimization device of claim 12 , wherein dynamically generating the conversation template further comprises:
dynamically determining temporal goal-specific intents of the conversation between the callee and the caller allocated to the callee in each of one or more timeframes associated with the conversation;
calculating a probability of achieving the objective of the call based on the temporal goal-specific intents of the conversation; and
dynamically generating the conversation template, based on the temporal goal-specific intents of the conversation and the probability of achieving the objective of the call.

14. The call distribution optimization device of claim 12 , wherein the ANN is a Long Short-Term Memory Network (LSTM), and wherein the ANN is trained with the temporal goal-specific intents of the conversation dynamically determined between each of the one or more callers and the set of callees allocated to each of the one or more callers.

15. The call distribution optimization device of claim 9 , wherein the AI model is an Asynchronous Actor Critic Network model.

16. A non-transitory computer-readable storage medium having stored thereon, a set of computer-executable instructions causing a computer comprising one or more proce

Assignees

• Wipro Ltd

Inventors

• Banerjee Sourish

Classifications

• H04M3/5233

  Operator skill based call distribution · CPC title

• H04M3/5158 Primary

  in combination with automated outdialling systems (devices for signalling identity of wanted subscriber H04M1/26) · CPC title

• G10L15/16

  using artificial neural networks · CPC title

• G10L15/1815

  Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning · CPC title

• G10L15/22

  Procedures used during a speech recognition process, e.g. man-machine dialogue · CPC title