Behavioral Pairing for Case Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current case assignment strategies in customer service centers, such as FIFO and management-based assignments, often lack confidence and relevant information, leading to suboptimal case allocations and performance issues.
Innovation Solution
Implementing a method and system that uses behavioral pairing to reassess and reassign cases, combining management expertise with artificial intelligence and big data analysis to optimize case assignments, allowing for both collaborative and non-collaborative approaches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If FIFO or management-based assignment strategies are used, then case allocations can be made quickly and simply, but the assignment performance and outcomes are suboptimal due to lack of confidence and relevant information
Solution Approach 1:
The patent introduces an automated case assignment system that acts as an intermediary between cases and agents. This system uses behavioral pairing algorithms to analyze historical data, agent characteristics, and case requirements, thereby mediating the assignment process to achieve both speed and reliability. The system processes assignments automatically while incorporating sophisticated analysis that management-based approaches cannot perform at scale.
Solution Approach 2:
The patent replaces manual management-based assignment mechanisms with an automated computational system. Instead of relying on human managers to evaluate and assign cases based on limited information, the system uses algorithms that process large datasets including agent performance history, case complexity metrics, and outcome predictions, thereby substituting mechanical human decision-making with a more reliable automated process.
2Ease of operation
If management assigns cases based on available information, then assignments can be made with some rationale, but confidence is low due to lack of relevant information
Solution Approach 1:
The patent performs preliminary analysis and data gathering before case assignments are made. The system pre-processes agent performance data, case characteristics, and historical outcomes to create ready-to-use assignment recommendations. This preliminary action ensures that when assignments are needed, the system already has processed information and confidence scores, eliminating the need for managers to gather information manually at the moment of assignment.
Solution Approach 2:
The patent implements feedback loops where assignment outcomes are continuously monitored and fed back into the system. This feedback mechanism allows the system to learn from past assignments, refine its behavioral pairing algorithms, and improve confidence in future recommendations. The system tracks which assignment strategies produce best results and adjusts its recommendations accordingly, creating a self-improving loop that increases measurement precision over time.
3Device complexity
If traditional assignment strategies are used, then the system remains simple to operate, but suboptimal performance is achieved in subrogation recoveries, patient care, and debt collection
Solution Approach 1:
The patent segments the case assignment process into distinct functional modules: data collection, behavioral pairing analysis, recommendation generation, assignment execution, and outcome tracking. Each module handles a specific aspect of the assignment process, allowing the complex system to be managed through modular components. This segmentation enables sophisticated performance optimization while maintaining operational simplicity through clear separation of concerns.
Data Source
AI summary
Techniques for case allocation are disclosed. In one particular embodiment, the techniques may be realized as a method for case allocation comprising receiving, by at least one computer processor, at least one case allocation allocated using a first pairing strategy, and then reassigning, by the at least one computer processor, the at least one case allocation using behavioral pairing.