As IT service desks handle growing numbers of support requests, manually sorting and routing tickets can take valuable time. MAASTRO wanted to streamline this process and help its IT team get incoming requests to the right place faster.
Maastro, a non-profit organization dedicated to providing optimal oncological treatments, manages a substantial IT support workload, addressing up to 1,000 inquiries monthly. These tickets range from reporting hardware issues to modifying electronic health records. The frontline team was responsible for assigning each ticket to the appropriate department, a process that was both complex and time-consuming. To enhance efficiency, Maastro sought an AI-driven solution to automate ticket classification.
1
Automated classification of IT support tickets.
2
Achieved an accuracy rate of 80%-85% in ticket categorization.
3
Continuous improvement through regular model retraining.
Team Raccoons collaborated with Maastro to develop an automated system designed to classify tickets into various manually selected categories. The pilot phase involved collecting data from the past five years, cleaning irrelevant information, and training the model exclusively on categories with sufficient data. The trained model was tested on unseen data, achieving an impressive accuracy rate of 80%-85% with exceptional speed.
In the production environment, a human-in-the-loop approach was implemented. The model assigns a probability score to each ticket, indicating its likelihood of belonging to a specific category, and produces a top-three list of categories. If the model is uncertain (e.g., a ticket scores more than 50% or less than 10% in all categories), the ticket is escalated to the first line for human intervention.
The final model has been operational at Maastro since May 2024. To ensure continuous improvement and optimal performance, the model is retrained every six months, evaluating performance across categories and incorporating new categories as needed. This ongoing collaboration ensures the AI-driven ticket classification system remains up-to-date and effective.
.jpg)
%20copy.webp)
