Action Classifier Filtering for Contact Center Knowledge Base Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing contact centers face inefficiencies in managing knowledge base queries during customer interactions, leading to increased costs and suboptimal service quality due to unnecessary searches and manual intervention.
Innovation Solution
A method involving an action classifier model is used to selectively initiate knowledge base queries based on customer actions, distinguishing between actions requiring and not requiring a search, thereby optimizing query initiation during ongoing conversations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If knowledge base searches are initiated for all customer actions, then service quality is maintained through comprehensive information retrieval, but system efficiency deteriorates due to unnecessary searches and increased costs
Solution Approach 1:
The system performs preliminary classification of customer actions into categories before initiating knowledge base searches. The action classifier model预先 determines whether a search is necessary based on the customer action type, preventing unnecessary searches from occurring in the first place. This resolves the contradiction by maintaining reliability for actions that need searches while improving productivity by eliminating wasteful searches.
Solution Approach 2:
The action classifier model serves as an intermediary between customer actions and knowledge base searches. It acts as a filtering layer that decides whether to forward a query to the knowledge base, thereby maintaining service quality for relevant queries while blocking unnecessary searches that would reduce system efficiency and increase costs.
2Productivity
If knowledge base searches are selectively initiated based on customer action classification, then system efficiency is improved by reducing unnecessary searches, but service quality may deteriorate if relevant searches are missed
Solution Approach 1:
The system performs preliminary classification of customer actions into categories before initiating knowledge base searches. The action classifier model预先 determines whether a search is necessary based on the customer action type, preventing unnecessary searches from occurring in the first place. This resolves the contradiction by maintaining reliability for actions that need searches while improving productivity by eliminating wasteful searches.
Solution Approach 2:
The system incorporates feedback mechanisms where classification data is continuously refined based on search outcomes and service results. This feedback loop ensures that the action classifier model learns from past performance, improving its accuracy in identifying which actions require searches, thereby maintaining service quality while optimizing system efficiency.
3Reliability
If manual intervention is used to determine when to search the knowledge base, then service quality is maintained through human judgment, but costs increase and productivity decreases
Solution Approach 1:
The system implements self-service through the action classifier model, which automatically determines whether knowledge base searches are needed based on customer action classification. This eliminates the need for manual human judgment in each interaction, reducing operational complexity and costs while maintaining service quality through automated, consistent decision-making.
Solution Approach 2:
The patent replaces the mechanical system of manual human intervention with an automated machine learning-based classification system. The action classifier model substitutes human agents in the decision-making process, automatically determining when searches are necessary based on learned patterns from historical data, thereby reducing operational complexity while maintaining or improving service quality.
Data Source
AI summary
A method in a contact center for generating an action classifier model and use thereof in selectively initiating turn set queries of a knowledge base to assist agents in real time during ongoing conversations with customers. The method includes: generating an action classifier model; receiving classification data that classifies a first plurality of the customer actions found in training samples as belonging to a first action category for which a knowledge base search is deemed needed, and a second plurality of the customer actions as belonging to a second action category for which a knowledge base search is deemed not needed; and using the action classifier model and the received classification data to perform a query filtering routine for selectively initiating a turn set query for a present turn set occurring in an ongoing conversation between an agent and customer.