Interaction Categorization Visualization Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing interaction analysis systems in organizations, such as call centers, struggle to effectively extract and visualize valuable data from categorized interactions, limiting their ability to understand complex issues and optimize business processes.
Innovation Solution
A method and apparatus for visualizing interaction categorization, which involves receiving interactions, defining categories and criteria, categorizing interactions, determining category networks, and extracting key-phrases to create a network model that represents connections between categories and key-phrases, allowing for deeper understanding and insight extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional interaction categorization methods are used, then interactions can be classified into categories, but the ability to extract and visualize valuable data from the categorization is limited
Solution Approach 1:
The patent combines multiple data extraction functions (transcription, word spotting, emotion detection, talkover analysis) into a unified categorization system that automatically generates structured information. This merging of functions enables comprehensive data extraction while managing system complexity through integrated processing.
Solution Approach 2:
The patent introduces an intermediary visualization layer that transforms raw categorization data into meaningful network representations. This intermediary system includes components for determining category networks, key-phrase networks, and relationships, acting as a mediator between raw data and actionable insights.
2Measurement precision
If comprehensive categorization of interactions is performed, then structured information is obtained, but the capability to derive deeper insights and patterns is insufficient
Solution Approach 1:
The patent segments the analysis process into distinct components: interaction categorization, key-phrase extraction, network determination, and relationship analysis. This segmentation allows each component to specialize in specific tasks, improving measurement precision while maintaining productivity through modular processing.
Solution Approach 2:
The patent transitions from traditional flat categorization to multi-dimensional network representations, adding dimensions of relationships, key-phrases, and interactions between categories. This dimensional expansion enables deeper insight generation while maintaining efficiency through systematic network analysis.
3Loss of information
If manual analysis of categorized interactions is used, then detailed insights can be obtained, but the process is time-consuming and inefficient
Solution Approach 1:
The patent implements self-service automation where the system automatically performs transcription, word spotting, emotion detection, talkover analysis, and network determination without requiring manual intervention. This automation maintains comprehensive data extraction while eliminating time-consuming manual processes.
Solution Approach 2:
The patent performs preliminary automated processing of interactions including transcription, categorization, and key-phrase extraction before analysis is needed. This preliminary action prepares data in advance, reducing the time required for actual insight generation while maintaining extraction completeness.
Data Source
AI summary
A method and apparatus for visualization of call categorization, comprising steps and components for: defining or receiving a definition for one or more categories and criteria for each category; receiving or capturing interactions; categorizing the interactions into the categories; determining relations between the categories; determining layout for the categories and relations; and visualizing the layout. The method and apparatus can further comprise steps and components for extracting key-phrases, determining connections between key-phrases, connections between categories based on key-phrases, and connections between categories and key-phrases, and visualizing the categories, key-phrases and connections. The method and apparatus can further comprise steps and components for training models upon which the relations between categories and relations between key-phrases are determined.