Distributed Contact Center Routing with Pre-computed Target Lists
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current contact center routing systems are inflexible, complex, and often rely on a single central routing engine, which limits their ability to handle large-scale, distributed routing decisions and fails to optimize interactions effectively in complex environments.
Innovation Solution
A distributed routing system that generates a list of all potential targets for an interaction without considering their current state, allowing for parallel processing and global optimization, and uses fitness parameters to select the best target based on various criteria such as skills, availability, and historical performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single central routing engine is used, then routing decisions can be made centrally, but the system cannot handle large-scale distributed routing decisions effectively
Solution Approach 1:
The patent divides the centralized routing engine into multiple distributed routing engines across different servers. Each routing engine can independently generate target lists and make routing decisions for specific interaction types or geographic regions, enabling the system to handle large-scale distributed routing decisions while maintaining centralized coordination through shared databases and communication protocols.
2Measurement precision
If the routing system considers current state of all potential targets, then routing accuracy improves, but processing time and system complexity increase
Solution Approach 1:
The patent implements pre-computation of target lists by generating lists of potential targets before interactions arrive. The system pre-loads agent skills, historical performance data, and other target attributes into cached structures, so that when an interaction needs routing, the system can quickly filter and rank pre-prepared targets rather than gathering all data in real-time, significantly reducing routing decision time while maintaining accuracy.
3Ease of manufacture
If skills are assigned administratively based on training completed, then assignment is simple, but the skills do not reflect actual capabilities of agents
Solution Approach 1:
The patent implements a feedback mechanism where the system continuously monitors actual agent performance on interactions and uses this data to refine skill assignments. Historical performance data, customer satisfaction scores, and interaction outcomes are fed back into the routing engine to adjust and update agent skill profiles over time, ensuring that skill assignments reflect actual capabilities rather than just administrative assignments based on training completion.
4Ease of operation
If basic skills-based routing is used, then implementation is straightforward, but the system cannot handle the complex needs of larger contact centers
Solution Approach 1:
The patent transforms static skills-based routing into a dynamic system where routing criteria can be adjusted in real-time based on changing conditions. The system allows dynamic modification of routing rules, priority scores, and target selection algorithms in response to varying interaction types, agent availability, and business priorities, enabling the system to handle complex needs of large contact centers while maintaining ease of operation through a unified routing framework.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system for optimized and distributed routing of interactions, comprising a plurality of media servers operating adapted to handle interactions of one or more specific media types, a statistics server, a routing database, and a routing server. Upon receiving or initiating an interaction of a specific media type, a first media server from the media servers sends a route request message comprising at least an interaction identifier and a collection of interaction-specific data to the routing server, the statistics server receives event notifications from the media servers and computes one or more statistics, and the routing server executes, using statistical data from the statistics server and data from the routing database, a routing script in response to the route request message and sends a response to the first media server directing it to route the interaction to a specific target resource.