Automated Problem Statement Extraction from Support Transcripts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Customer support agents spend significant time summarizing calls, and the quality and style of summaries vary significantly between agents, making it inefficient and inconsistent.
Innovation Solution
A method and system that use trained classification and regression models to identify and select problem-relevant sentences in customer support call transcripts by predicting regions containing problem statements, filtering out uninformative content, and estimating the relevance of customer utterances within these regions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If agents manually summarize customer support calls, then summaries can be created, but agents spend significant time on this task and quality varies significantly between agents
Solution Approach 1:
The system enables automated self-service summarization by using machine learning models to automatically identify and extract problem-relevant sentences from call transcripts, eliminating the need for manual agent summarization while maintaining consistent quality across all summaries
Solution Approach 2:
The patent replaces the mechanical manual summarization process with an automated computational system that uses trained classification and regression models to analyze transcripts and extract relevant problem statements, substituting human cognitive effort with algorithmic processing
2Loss of time
If automated models are used to extract problem statements, then time is saved and consistency is improved, but system complexity increases
Solution Approach 1:
The automated summarization system is divided into distinct functional modules: a classification model that identifies customer utterances, a regression model that estimates relevancy scores, and a selection mechanism that extracts problem statements. This segmentation allows each component to be independently trained, optimized, and maintained, managing overall system complexity
3Adaptability or versatility
If manual summarization is used, then flexibility in handling diverse call types is maintained, but significant time is spent on each summary
Solution Approach 1:
The system adapts to diverse call types by dynamically adjusting parameters such as relevancy thresholds and model confidence levels. The regression model estimates relevancy scores that can be tuned based on different call categories, allowing the same automated system to handle varied support scenarios efficiently while maintaining adaptability
Data Source
AI summary
Systems and methods for automatically identifying problem-relevant sentences in a transcript are disclosed. In an example method, a transcript may be received of a first support call. A region of the first support call transcript may be identified, and first customer utterances may be detected in the region using a trained classification model. A trained regression model may estimate a relevancy to the problem statement of each of the first customer utterances, and one or more most problem-relevant statements may be selected from the first customer utterances, based on the estimated relevancies.