Feedback-Based Communication Router for Contact Centers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing contact center routing systems for text-based communications often fail to accurately route communications to the appropriate resources due to outdated predefined rules, requiring frequent updates and not leveraging feedback effectively.
Innovation Solution
Implement a system where inbound communications are classified based on feedback from prior interactions, allowing for dynamic routing to different contact center queues based on agent feedback, user feedback, and historical communication data, with the use of a communication classifier that learns from feedback to generate new routing rules over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If predefined rules are used for routing text-based communications, then routing can be implemented with basic infrastructure, but routing accuracy deteriorates and frequent rule updates are required
Solution Approach 1:
The system implements feedback loops where routing outcomes are continuously monitored and fed back to the machine learning model. The model learns from actual routing results, agent performance data, and communication outcomes to automatically refine routing decisions, eliminating the need for manual rule updates while improving accuracy over time.
Solution Approach 2:
The routing system performs self-optimization through the machine learning model that automatically adjusts routing parameters and decisions based on learned patterns from historical data. The system serves itself by continuously improving without external intervention, replacing manual rule maintenance with autonomous adaptation.
2Reliability
If feedback-based dynamic routing is implemented, then routing accuracy improves, but system complexity increases
Solution Approach 1:
The machine learning model serves multiple functions simultaneously: it classifies communications, determines routing decisions, learns from feedback, and adapts to changing patterns. This multi-functionality consolidates what would otherwise require separate systems for classification, decision-making, and optimization into a single unified component.
Solution Approach 2:
The machine learning model acts as an intermediary layer between raw communication data and routing execution. It processes input communications, applies learned patterns, and outputs routing decisions, mediating between the complexity of data analysis and the simplicity of routing execution while managing system complexity.
3Adaptability or versatility
If traditional SBR with predefined rules is used, then system implementation is straightforward, but adaptability to new communication patterns deteriorates
Solution Approach 1:
The routing system transitions from static predefined rules to dynamic machine learning-based decisions. The model continuously adapts to new communication patterns, agent availability, and routing outcomes, making the system flexible and responsive to changing conditions rather than rigid and fixed.
Solution Approach 2:
The system changes key parameters from fixed rule definitions to learned probability distributions and decision boundaries. The machine learning model adjusts routing parameters dynamically based on input features and historical patterns, enabling adaptation to new communication types and patterns without structural changes.
Data Source
AI summary
An inbound communication is received by a contact center. For example, the inbound communication may be a text-based communication. The inbound communication is classified based on feedback from one or more prior inbound communications. For example, the feedback may be based on whether a previous communication was transferred, was dropped, based on feedback provided by a contact center agent, based on feedback provided by a user, and/or the like. In response to classifying the first inbound communication, the inbound communication is routed differently in the contact center. For example, the inbound communication is routed to a different contact center queue based on the feedback of the previous communication.