Memory Neural Network Agent Scoring from Contact Center Call Summaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Evaluating contact center agent performance is labor-intensive, subjective, and resource-intensive, with traditional methods introducing sampling biases due to the sheer volume of interactions, hindering scalability and consistency in evaluations.
Innovation Solution
A system using memory neural networks analyzes call recordings to generate conversation summaries, determine agent performance categories, and provide cumulative performance analytics, leveraging AI/ML to monitor and score agent interactions efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional agent performance analysis techniques are used, then resource consumption is reduced, but analysis completeness deteriorates due to the sheer volume of interactions
Solution Approach 1:
The patent replaces manual supervisor-led performance analysis with an automated memory neural network system that processes call recordings. The neural network automatically extracts performance metrics, generates summaries, and evaluates agent performance against predefined criteria, eliminating the need for human supervisors to manually review each interaction while achieving comprehensive analysis of all calls.
Solution Approach 2:
The system creates a simplified representation of each call interaction through conversation summaries generated by the memory neural network. These summaries capture essential performance information without requiring analysis of the complete original recordings, allowing comprehensive evaluation of all interactions while reducing processing resources needed for each individual analysis.
2Loss of time
If traditional supervisor-led agent performance analysis is used, then evaluation subjectivity is reduced, but time consumption increases
Solution Approach 1:
The patent implements automated performance evaluation using a memory neural network that processes call recordings and generates performance assessments without human intervention. The system automatically compares agent performance against predefined criteria, generates summaries, and identifies performance categories, completely replacing the manual supervisor-led evaluation process and eliminating time consumption associated with human review.
3Reliability
If comprehensive analysis of all interactions is performed, then evaluation reliability is improved, but system complexity increases
Solution Approach 1:
The patent divides the complex task of comprehensive performance analysis into manageable segments through the memory neural network architecture. The system processes calls individually, generates discrete conversation summaries for each, evaluates them against predefined performance criteria, and aggregates results. This segmentation allows reliable analysis of all interactions while keeping each processing unit simple and manageable.
Solution Approach 2:
The system transforms complex call recordings into simplified performance parameters through the memory neural network. By converting raw audio data into structured conversation summaries with extracted performance metrics, the system reduces complexity while maintaining comprehensive analysis capability. The predefined performance categories and scoring thresholds provide clear, objective criteria that simplify the evaluation process while ensuring reliable results.
4Quantity of substance
If sampling of interactions is used, then resource consumption is reduced, but data representativeness deteriorates
Solution Approach 1:
The system creates a complete copy of all interactions through conversation summaries generated by the memory neural network. Instead of analyzing subsets of calls, the system processes every call recording and generates a summary representation, ensuring that all interactions are represented in the final performance evaluation. This complete copying approach maintains data representativeness while reducing the complexity of processing through summarization.
Data Source
AI summary
A system for complexity assessments for agent performance analysis using a memory neural network according to an embodiment includes receiving a plurality of call recordings of calls between a contact center agent and one or more users, generating, for each call recording, a respective conversation summary of the call recording, analyzing each respective conversation summary using the memory neural network to determine a respective agent performance category for performance of the agent during the respective call, each respective agent performance category being selected from a plurality of predefined agent performance categories and associated with a respective call score, comparing each respective call score to a predefined threshold value and updating a topics array based on the respective conversation summary in response to determining that the respective call score is below the predefined threshold value, and determining a total performance score for the agent based on each respective call score.