Automated Intent Classification via Action-Utterance Correlation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Building and maintaining intent classifiers in customer care systems is a time-consuming and error-prone manual process, making it inefficient for accurately determining user intent from conversations.

Innovation Solution

An automated intent classification system that identifies user intent by correlating actions with utterances, determining intent based on these correlations, and associating specific actions with the intent for future identification, thereby streamlining the process and improving accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual processes are used to build and maintain intent classifiers, then flexibility and control are maintained, but the process becomes time-consuming and error-prone

Engineering Contradiction:

Improveaccuracy of intent classificationVSAvoidtime to build and maintain classifiers

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system enables self-service automation where the intent classification system automatically determines intents by correlating actions with utterances, eliminating the need for manual classifier construction and maintenance while improving accuracy and reducing time investment

Inventive Principle:

Principle #25Self-service

2Productivity

If automated systems are implemented for intent classification, then efficiency and speed improve, but system complexity increases

Engineering Contradiction:

Improvespeed of intent identificationVSAvoidcomplexity of classification system

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The system introduces an intermediary mechanism that automatically correlates actions with utterances to determine intent, simplifying the overall system architecture by using observable action-utterance relationships rather than requiring complex manual classifier construction

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system implements feedback loops where intent classifications are continuously refined based on the correlation between user utterances and system actions, allowing the system to automatically improve accuracy without increasing operational complexity

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS10068567B1System, method, and computer program for automatic management of intent classification

Publication Date: 2018.09.04 AMDOCS DEV LTD

AI summary

A system, method, and computer program product are provided for automatic management of intent classification. In operation, an intent classification system identifies a conversation including a plurality of utterances for which to identify intent of a user associated with the plurality of utterances. The intent classification system identifies one or more actions performed in association with the plurality of utterances. Further, the intent classification system correlates at least one of the one or more actions to at least one of the plurality of utterances based on which of the one or more actions were performed as a result of the plurality of utterances. The intent classification system automatically determines an intent of the user based on the correlation of the at least one of the one or more actions to the at least one of the plurality of utterances. Additionally, the intent classification system stores an indication of the intent based on the at least one of the plurality of utterances. Moreover, the intent classification system associates one or more specific actions with the indication of intent such that subsequent utterances are capable of being used by the intent classification system to automatically identify an intent of one or more users and to automatically identify specific actions that should be performed based on the automatically identified intent of the one or more users.