Ticket Prioritization Using Time-Based Machine Learning Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for prioritizing customer service tickets are inefficient due to low precision and low coverage, often resulting in missed events and false alerts, as they rely on rule-based systems that fail to accurately assess the likelihood of a ticket becoming distressed.
Innovation Solution
A method that involves obtaining customer service tickets, collecting dynamic and textual features, and applying them to multiple machine learning models trained on historical data corresponding to different time periods to generate a distress score, which indicates the likelihood of a ticket becoming distressed, thereby improving prioritization accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If rule-based systems are used for ticket prioritization, then the system is simple to implement, but the precision and coverage of prioritization are low
Solution Approach 1:
The patent segments the prioritization task by creating multiple machine learning models, each specialized for a specific time period (hourly, daily, weekly, monthly). This segmentation allows each model to focus on patterns relevant to its time window, improving overall prioritization accuracy while managing complexity through modular architecture.
Solution Approach 2:
The patent changes the approach from static rule-based parameters to dynamic machine learning models that are trained on historical data and adjusted over time. The models incorporate multiple features including account information, product data, textual content, and dynamic activities, transforming the prioritization from fixed rules to adaptive parameter-based predictions.
2Measurement precision
If multiple machine learning models are used for ticket prioritization, then the prioritization accuracy is improved, but the computational resources and time required increase
Solution Approach 1:
The computational workload is segmented across multiple specialized models, each handling a specific time period. This allows the system to process tickets efficiently by assigning them to the appropriate time-based model rather than requiring all models to process every ticket, reducing overall computational resource consumption while maintaining high accuracy.
Solution Approach 2:
The system applies machine learning models selectively based on the ticket age and required analysis depth. For newer tickets, more detailed multi-model analysis is applied, while older tickets may use simplified evaluation, optimizing the balance between accuracy and computational resource usage.
3Quantity of substance
If dynamic features are collected and processed for all tickets, then the prioritization coverage is improved, but the processing time and complexity increase
Solution Approach 1:
The patent segments the feature collection and processing based on time periods. Each machine learning model is trained on historical data from its specific time window and applies relevant features accordingly. This segmentation allows comprehensive feature analysis for tickets that need it while avoiding unnecessary processing for tickets that can be handled with simpler evaluation, reducing overall processing time.
Data Source
AI summary
Techniques are provided for customer service ticket prioritization using multiple time-based machine learning models. One method comprises obtaining a customer service ticket; collecting, in response to the obtaining, features related to the customer service ticket comprising dynamic information regarding activities performed to address the customer service ticket and textual information describing the customer service ticket; applying the collected features to one of multiple machine learning models to obtain a distress score indicating a likelihood that the customer service ticket will become a distressed customer service ticket, wherein the multiple machine learning models each correspond to different time periods and are trained on historical data and the collected features are applied to the one machine learning model based on an age of the customer service ticket; and processing multiple customer support tickets based on the distress score. The machine learning models are trained using cumulative corresponding historical training data associated with a respective one of hourly, daily, weekly and monthly time periods.