Automated Conversation Analysis for Contact Center Trend Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems for analyzing conversations in contact centers rely on manual data collection and analysis, which is time-consuming and delays the identification of trends and issues, and struggles to categorize conversations with phrases not previously identified.
Innovation Solution
A method that automatically filters, clusters, and names informative sentences from interactions to identify concepts and categorize conversations without human assistance, using saliency computation and term frequency-inverse document frequency analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual analysis of conversations is used, then categorization accuracy can be maintained, but analysis time and labor requirements increase significantly
Solution Approach 1:
The system performs automatic self-analysis of conversations using computational algorithms. The conversation analysis system independently processes, categorizes, and identifies patterns in conversations without requiring manual human intervention, thereby maintaining accuracy while significantly reducing analysis time and labor requirements
Solution Approach 2:
The patent replaces manual mechanical analysis with automated computational systems. The system uses algorithms, natural language processing, and pattern recognition techniques to substitute human analysts, achieving both high accuracy and rapid processing speeds that manual methods cannot match
2Reliability
If predefined keyword-based categorization is used, then existing categories can be identified, but new or emerging topics cannot be detected
Solution Approach 1:
The system dynamically adapts its categorization capabilities by continuously learning from new conversation data. Instead of relying on static predefined keywords, the system evolves its understanding of topics and categories through automated pattern recognition, enabling it to detect and categorize emerging topics that were not previously in its knowledge base
Solution Approach 2:
The system incorporates feedback loops where analyzed conversations are used to refine and update categorization models. This feedback mechanism allows the system to learn from new patterns and improve its ability to detect both existing and emerging topics, maintaining reliability while enhancing adaptability to new conversational themes
3Measurement precision
If Bayesian networks with human-defined parameters are used, then correlation analysis can be performed, but system complexity and development time increase
Solution Approach 1:
The system automatically generates and refines its own correlation models without requiring human expertise in Bayesian network construction. The computational system independently analyzes conversation data, identifies relationships between topics and categories, and builds structured knowledge representations, thereby reducing system complexity while maintaining precise correlation analysis capabilities
Solution Approach 2:
The system performs preliminary automated learning and model building during the data processing phase. By pre-computing relationships and patterns from the data itself rather than requiring pre-defined human parameters, the system reduces the complexity of subsequent analysis while maintaining high precision in correlation detection
Data Source
AI summary
A method for identifying concepts in a plurality of interactions includes: filtering, on a processor, the interactions based on intervals; creating, on the processor, a plurality of sentences from the filtered interactions; computing, on the processor, a saliency of each the sentences; pruning away, on the processor, sentences with low saliency for generating a set of informative sentences; clustering, on the processor, the sentences of the set of informative sentences for generating a plurality of sentence clusters, each of the clusters corresponding to a concept of the concepts; computing, on the processor, a saliency of each of the clusters; and naming, on the processor, each of the clusters.