Complexity-Aware Call Routing in Heterogeneous Contact Centers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Call centers face challenges in efficiently routing incoming calls to either real or virtual agents, as existing methods fail to accurately assess the complexity of calls at the outset, leading to potential frustration for customers and inefficiencies in resource allocation.
Innovation Solution
A system and method that utilize natural language processing, spectral methods, and logistic regression to predict the class and complexity of a call based on the first utterance, employing a multi-armed parametric bandit formulation to dynamically route calls to either human or virtual agents, considering the cost and success rates of each type of agent.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If calls are routed to virtual agents without accurate complexity assessment, then automation extent increases, but call handling reliability deteriorates
Solution Approach 1:
The system performs preliminary complexity assessment of incoming calls using natural language processing and spectral methods before routing to virtual or human agents. This advance analysis predicts dialogue complexity based on the first customer utterance, enabling proactive routing decisions that prevent virtual agent failures before they occur.
Solution Approach 2:
The patent replaces manual assessment of call complexity with automated computational methods including natural language processing, spectral analysis, and machine learning models. This substitution enables scalable, consistent, and objective complexity evaluation that can handle high call volumes while maintaining accurate routing decisions.
2Reliability
If complex analysis is performed to detect when virtual agents face difficulty, then call handling reliability improves, but device complexity increases
Solution Approach 1:
Instead of performing complex analysis during the dialogue to detect difficulties, the system performs the complexity assessment in advance using the first customer utterance. This preliminary action eliminates the need for continuous monitoring and complex real-time analysis, reducing device complexity while maintaining high reliability.
Solution Approach 2:
The patent extracts only the most critical features from the first customer utterance to predict dialogue complexity, rather than analyzing the entire dialogue. This extraction approach reduces computational complexity while maintaining accurate predictions, avoiding the need for continuous complex analysis throughout the call.
3Reliability
If calls are routed to human agents directly without virtual agent attempt, then call handling reliability improves, but productivity decreases
Solution Approach 1:
The system applies different routing strategies based on the local characteristics of each call's complexity. Low-complexity calls are routed to virtual agents for automated handling, while high-complexity calls are directed to human agents. This localized quality approach ensures each call receives appropriate handling based on its specific needs, optimizing both reliability and productivity.
Solution Approach 2:
The patent changes the routing parameter from a static rule-based system to a dynamic complexity-based system. By varying the routing decision based on predicted dialogue complexity, the system can flexibly allocate calls to the most appropriate agent type, improving overall system productivity while maintaining high reliability for complex calls.
4Reliability
If more sophisticated routing strategies are implemented, then call handling reliability improves, but device complexity increases
Solution Approach 1:
The patent replaces complex rule-based routing logic with machine learning models that automatically learn optimal routing strategies from data. This substitution enables sophisticated routing decisions based on multiple factors including dialogue complexity, agent availability, and historical performance, without requiring manual configuration of complex rules.
Solution Approach 2:
The routing system continuously learns and adapts from call outcomes and complexity assessments, improving its routing decisions over time without external intervention. This self-service capability allows the system to automatically optimize routing strategies based on accumulated experience, enhancing reliability while maintaining manageable complexity.
Data Source
AI summary
A method for routing calls suited to use in a call center includes receiving a call from a customer, extracting features from an utterance of the call, and, based on the extracted features, predicting a class and a complexity of a dialog to be conducted between the customer and an agent. With a routing model, a routing strategy is generated for steering the call to one of a plurality of types of agent (such as to a human or a virtual agent), based on the predicted class and complexity of the dialog and a cost assigned to the type of agent. A first of the plurality of types of agent is assigned a higher cost than a second of the types of agent. The routing strategy is output.