Virtual Agent Training via Fallback Utterance Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for training virtual agents are time-consuming and prone to errors, requiring manual identification of new intents and training phrases, which can lead to inaccurate responses.
Innovation Solution
A method and system that utilize fallback analysis to classify and cluster utterances, generating new intent categories and training phrases through machine learning models, automating the process of identifying and recommending new intents based on fallback logs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual identification of new intents and training phrases is used, then virtual agent training can be performed, but the process is time-consuming and error-prone
Solution Approach 1:
The system automatically analyzes fallback logs, clusters similar utterances, and generates training phrases and new intent categories without human intervention. The virtual agent training process serves itself by extracting learning data from its own failure cases (fallback logs), eliminating the need for manual identification of training data while improving both accuracy and reducing time consumption.
2Adaptability or versatility
If manual identification of new intents is performed, then new intent categories can be created, but the process is labor-intensive and may result in poor intent training
Solution Approach 1:
The manual mechanical process of identifying and categorizing intents is replaced with an automated computational system. The system uses machine learning models to analyze fallback logs, automatically cluster similar utterances, and generate intent categories, substituting human cognitive work with algorithmic processing that is both more efficient and more scalable.
Solution Approach 2:
The system introduces an intermediary processing layer between user queries and intent classification. This intermediary automatically analyzes fallback logs, identifies patterns through clustering, and generates training data, serving as a mediator that transforms raw fallback data into structured training phrases and intent categories without direct human involvement.
3Productivity
If traditional training methods are used, then virtual agent can be trained, but manual effort and potential for errors increase
Solution Approach 1:
The system implements a feedback loop where fallback logs (representing training failures) are continuously analyzed and fed back into the training process. By automatically clustering fallback utterances and generating new training phrases from these failures, the system converts errors into learning opportunities, continuously improving accuracy while maintaining high productivity through automation.
Data Source
AI summary
A method and system for training a virtual agent through fallback analysis is provided herein. The method comprises obtaining a plurality of fallback utterances. The method further comprises classifying the plurality of fallback utterances into one or more of existing intent categories, via a Machine Learning (ML) model. The method further comprises upon unsuccessful classification of one or more utterances of the plurality of fallback utterances, clustering the one or more utterances into one or more groups based on similarities among the one or more utterances, via the ML model. Further, the method comprises generating labels for the one or more groups to determine names of new intent categories associated with the one or more utterances.