Model Stacking for Direct Support Ticket Routing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing support ticket handling systems require all tickets to be evaluated by primary support personnel before being reassigned to secondary support teams, leading to inefficiencies and resource wastage.
Innovation Solution
Implementing a model stacking technique that utilizes machine learning models of varying complexities to categorize support tickets, allowing direct routing to appropriate support groups based on trained neural networks and database schemas, thereby bypassing initial evaluation by primary support.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all support tickets are evaluated by primary support personnel before reassignment, then tickets are thoroughly assessed, but primary support staff workload increases and resources are wasted on complex tickets
Solution Approach 1:
The patent applies preliminary action by implementing an automated triage system that evaluates support tickets before they reach primary support staff. Machine learning models analyze ticket content, historical data, and complexity metrics in advance to determine appropriate routing, so that primary support personnel only handle tickets they are actually capable of resolving. This preliminary classification prevents waste of primary support resources on complex tickets that require secondary support expertise.
Solution Approach 2:
The patent introduces an intermediary automated triage system between ticket submission and primary support evaluation. This intermediary uses machine learning models to act as a mediator that pre-assesses tickets, determines complexity levels, and routes appropriate tickets directly to secondary support teams. This intermediary layer maintains assessment quality while protecting primary support staff from being overloaded with unsuitable tickets.
2Measurement precision
If machine learning models of varying complexities are stacked, then routing accuracy improves, but system complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the ticket routing system into multiple specialized machine learning models of varying complexities. Each model in the stack handles specific aspects of ticket evaluation (e.g., initial filtering, complexity assessment, routing decisions). This segmentation allows the system to achieve high categorization accuracy through ensemble voting while managing complexity by organizing models in a hierarchical stack rather than a monolithic structure.
Solution Approach 2:
The patent uses the composite materials principle by creating a composite routing system that combines multiple machine learning models with different strengths and complexities. Like composite materials that combine different substances to achieve superior properties, the model stack combines simpler and more complex ML models to achieve higher routing accuracy than any single model could provide alone, while the modular composite structure manages overall system complexity.
Data Source
AI summary
Example methods and systems are directed to categorizing support tickets for more efficient handling by support staff. Support staff may be divided into multiple support groups. An incoming support ticket is converted to a machine representation and provided as input to one or more trained machine learning models. Based on the output from the one or more trained machine learning models, the support ticket is routed to one of the support groups. As a result, some tickets will be directly routed to higher-level support groups instead of having all tickets first be evaluated by L1 support personnel. Accordingly, support staff resources are conserved. A model stacking technique may be used in which models of varying complexities, ranging from very simple to highly complex, are stacked in sequence one after another.