Agent Rating Prediction Routing for Contact Centers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing quality management systems in contact centers, such as Avaya Aura Workforce Optimization, do not effectively utilize agent ratings to route contacts to the best-suited agents due to biased sampling methods, leading to poor routing decisions and reduced productivity.
Innovation Solution
A mechanism for agent rating prediction and routing that employs predictive models based on contact and agent attributes, using machine learning techniques like regression and Bayesian models to maximize accuracy in assigning contacts to qualified agents, thereby optimizing contact-agent assignments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If random sampling of calls is used to rate agents, then quality management can be implemented with reduced supervision workload, but the sampled contacts become unrepresentative of the contacts an agent could possibly handle, introducing substantial bias into the sample
Solution Approach 1:
The system performs preliminary actions by collecting and storing attributes of all contacts in a context store before routing decisions are made. This allows the system to have all necessary information available in advance to make accurate routing predictions without needing to randomly sample contacts for rating, thereby eliminating bias while maintaining reduced supervision workload.
Solution Approach 2:
The patent introduces a predictive modeling system as an intermediary between contact attributes and routing decisions. This intermediary uses machine learning models to predict contact outcomes based on historical data and attributes, providing accurate rating information without requiring actual random sampling of contacts, thus resolving the contradiction between reduced workload and accurate measurement.
2Device complexity
If routing decisions are made based on biased sampling results, then the routing system operates with limited information processing requirements, but poor routing decisions are made leading to loss of productivity
Solution Approach 1:
The system performs preliminary actions by pre-collecting and storing comprehensive contact attributes, agent attributes, and historical outcome data in context stores before routing decisions are needed. This preliminary data collection enables the predictive modeling system to make accurate routing decisions without requiring complex real-time analysis, thus maintaining low information processing requirements while improving productivity through better routing decisions.
Solution Approach 2:
The patent transitions from traditional one-dimensional routing based on simple attributes to a multi-dimensional approach using predictive models that consider historical outcomes, contact attributes, agent attributes, and interaction patterns. This dimensional expansion enables more accurate routing predictions without proportionally increasing processing complexity, as the models are trained in advance on comprehensive datasets.
Data Source
AI summary
An agent rating prediction and routing mechanism provided by a contact center communication system for work assignment optimization is described along with various methods and mechanisms for administering the same. The prediction system proposed herein analyzes past agent performance, agent attributes, contact attributes, and customer attributes to calculate an outcome value and to provide a performance prediction for use in work item routing to contact center resources.