Conversational Transcript Analysis for Automated Knowledge Base Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Call centers face challenges in efficiently analyzing and adapting to new issues due to the limitations of manually authored knowledge bases, which are slow to update and require significant manpower, making it difficult to identify effective procedural sequences from conversational transcripts.
Innovation Solution
A method for analyzing conversational transcripts by clustering and representing them as sequences of sub-procedural text segments, using algorithms like CAARD to identify frequent and distinct procedural sequences, and updating knowledge bases with new information from unrecorded situations, enabling the automatic generation of procedural documents for both agents and customers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If manually authored knowledge bases are used, then agents can access structured information, but the system cannot quickly adapt to new problems and requires tremendous manpower for updates
Solution Approach 1:
The system automatically mines procedural sequences from historical conversational transcripts without requiring manual intervention. The automated knowledge discovery process enables the system to self-update with new problem-solution patterns, eliminating the need for manual knowledge base augmentation while rapidly adapting to new problems.
Solution Approach 2:
The patent replaces the manual mechanical process of knowledge base authoring with an automated computational system. Machine learning algorithms and text mining techniques automatically extract procedural sequences from unstructured conversational data, substituting human effort with automated information processing.
2Reliability
If manually authored knowledge bases are used, then agents can access structured information, but manual augmentation consumes tremendous manpower and may not be exhaustive
Solution Approach 1:
The system automatically discovers and adds procedural sequences from historical transcripts, ensuring comprehensive coverage of all problem types encountered in customer interactions. This self-updating mechanism ensures the knowledge base becomes progressively more exhaustive without additional manual effort.
Solution Approach 2:
The system performs preliminary analysis of historical conversational transcripts to pre-extract procedural sequences before they are needed. By proactively mining and storing knowledge patterns in advance, the system ensures comprehensive coverage is achieved automatically rather than through reactive manual updates.
3Ease of operation
If keyword search over manually authored knowledge bases is used, then agents can obtain useful information, but the utility is limited by the smartness of the agent to enter the right keywords
Solution Approach 1:
The patent replaces manual keyword searching with automated natural language processing. The system automatically analyzes conversational transcripts, identifies situations and problems, and retrieves relevant procedural sequences without requiring agents to formulate precise search keywords, thereby eliminating the dependency on agent skill while maintaining retrieval accuracy.
Solution Approach 2:
The system introduces an intermediary natural language processing layer between the agent's query and the knowledge base. This intermediary automatically interprets conversational input, identifies relevant situations, and retrieves appropriate procedural sequences, acting as a mediator that translates natural language into precise information retrieval without requiring manual keyword formulation.
4Productivity
If automated analysis of conversational transcripts is implemented, then procedural sequences can be automatically identified, but the complexity of analyzing unstructured data increases
Solution Approach 1:
The system segments the complex task of analyzing unstructured conversational transcripts into distinct modular components: speech recognition, natural language processing, situation identification, and procedural sequence extraction. Each module handles a specific aspect of the analysis, making the overall complex system manageable and maintainable while enabling rapid knowledge discovery.
Data Source
AI summary
Analyzing transcripts of conversation between at least two users by receiving input information from a first user via a voice call, creating conversational transcripts from the information received from the first user, selecting at least one defined situation from a list of defined situations, identifying the selected situation in the conversational transcripts, identifying a set of procedural sequences by comparing the at least one identified situation in the conversational transcripts with knowledge derived from a corpus of historical conversational transcripts; and providing the set of procedural sequences to the first user.