Dynamic FAQ Knowledge Base Update via AI Transcript Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing agent assist solutions in call centers rely on manually updated knowledge bases, which are time-consuming and often result in outdated information, reducing the effectiveness of the support provided to contact center agents.
Innovation Solution
A dynamic system that utilizes a computing system to analyze transcripts of contact center interactions, generate question-answer pairs using an artificial intelligence model, and update a frequently asked questions knowledge base in real-time by determining which question-answer pairs are frequently asked and approved by contact center agents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If manual updating of knowledge base is used, then implementation simplicity is maintained, but information freshness and accuracy deteriorate
Solution Approach 1:
The system automatically generates FAQ candidates from interaction transcripts and updates the knowledge base without requiring manual intervention. The AI model processes transcripts, identifies frequently asked questions, and autonomously maintains the knowledge base, making the system self-sufficient for knowledge updates.
Solution Approach 2:
The manual mechanical process of knowledge base updating is replaced with an automated AI-based system. The artificial intelligence model analyzes transcripts and generates FAQ entries, substituting human manual work with automated computational processes that continuously update the knowledge base.
2Productivity
If manual updating of knowledge base is used, then system complexity is reduced, but update frequency and productivity deteriorate
Solution Approach 1:
The system continuously processes interaction transcripts and updates the knowledge base in real-time or near-real-time. Rather than periodic manual updates, the automated system maintains continuous operation, analyzing new interactions and refreshing FAQ content as they occur, ensuring constant improvement of knowledge base currency.
Solution Approach 2:
The automated AI system replaces manual updating operations, enabling high-frequency updates that would be impractical for human operators. The system processes multiple transcripts and generates numerous FAQ candidates automatically, achieving productivity levels unattainable through manual processes.
3Reliability
If dynamic automated updating is implemented, then knowledge base currency is improved, but processing time and computational resources increase
Solution Approach 1:
The system processes a subset of transcripts or focuses on identifying only the most significant FAQ candidates rather than analyzing every possible detail. By prioritizing high-impact updates and using frequency thresholds, the system achieves reliable knowledge base currency while avoiding unnecessary computational expenditure on low-value processing.
Solution Approach 2:
The system adjusts processing parameters such as frequency thresholds, confidence levels, and transcript sampling rates to optimize the balance between accuracy and resource consumption. By dynamically tuning these parameters, the system maintains high information accuracy while adapting computational resource usage to current needs and system capacity.
4Measurement precision
If AI model analysis of transcripts is used, then FAQ generation accuracy is improved, but processing speed deteriorates
Solution Approach 1:
The system applies AI analysis selectively to transcripts that meet certain criteria or focuses on identifying only the most probable FAQ candidates. By processing a partial set of high-priority transcripts first and using confidence thresholds, the system achieves high accuracy for critical FAQs while maintaining overall processing speed through selective application of intensive analysis.
Data Source
AI summary
A method of dynamically updating a frequently asked questions knowledge base for agent assist according to an embodiment includes receiving, by a computing system, a transcript of a contact center interaction between a contact center agent and a client, analyzing, by the computing system, the transcript to generate a plurality of question-answer pairs based on an artificial intelligence model, determining, by the computing system, whether each question-answer pair of the plurality of question-answer pairs is stored in a dynamic frequently asked questions knowledge base, determining, by the computing system, whether a question-answer pair of the plurality of question-answer pairs is frequently asked in response to determining that the question-answer pair is not stored in the dynamic frequently asked questions knowledge base, and updating, by the computing system, the dynamic frequently asked questions knowledge base to include the question-answer pair in response to determining that the question-answer pair is frequently asked.