DATAGROUP: Ticket Categorization for DATAGROUP Service Desk
The fully automated categorization of tickets with the machine learning based Almato Classifier brings significant savings for one of the largest IT service desks in Germany.
The Key Facts at a Glance
- Reduced throughput times
- Improved service and higher customer satisfaction
- Machine Learning-based prediction model
The Challenge
DATAGROUP is one of the leading German IT service providers, with its IT Service Desk serving as the cornerstone of its portfolio. Acting as the central point of contact for users and IT employees across numerous customer organizations, the service desk offers first-level support, receives and processes customer reports, and escalates issues when necessary. Teams at the service desk continuously monitor the status of fault resolutions, service requests, and other inquiries, all of which adhere to ISO 20000 certification standards. The organization's primary objectives are to maintain agreed service levels, continually optimize processes and free up employees from routine tasks to focus on more demanding consulting and service duties.
Handling thousands of tickets daily, the initial step involves accurately classifying each incoming ticket. Depending on the customer, this classification involves assigning various categories such as impact, urgency, ticket category, service, and ticket type (e.g., service request). This classification dictates further ticket processing, including routing to the appropriate processing group and prioritization, making it a crucial element for the efficient organization of the IT service desk.
The time required for this classification can amount to several minutes, depending on the scope of the ticket content. In total, this task can take hundreds or thousands of working hours per year.
As part of its automation strategy, DATAGROUP Service Desk GmbH aims to implement a powerful and easy-to-integrate solution for the automated categorization of incoming tickets.
The Almato Solution
DATAGROUP utilizes the Almato Classifier for automated classification. The product offers a very easy-to-use and fast way to train a robust machine learning-based prediction model. The commissioning was carried out by DATAGROUP's system administration department. The ITSM tool used communicates with the classification service via a REST service.
The necessary training and test data could be imported in a simple form (CSV file) using a REST service. The subsequent model generation is fully automated using a proprietary method including hyperparameter tuning. Additional training data can be uploaded for continuous learning and the service is then retrained. From then on, the incoming ticket data is transferred from the ITSM tool to the Almato Classifier and analyzed by it.
The service then returns a so-called multi-label classification. The out-of-the-box service then identifies the area in which the ticket should be classified, who the correct contact person is and what priority the case has. The ticket is then forwarded to the correct processor accordingly and an opening email is sent to the ticket creator.
Thanks to the innovative solution, employees who were otherwise largely occupied with classification are freed from this work step and can go straight to solving the problem.
The Success
Ticket categorization by the Almato Classifier fully automates a monotonous routine process. The qualified experts at the IT Service Desk can therefore concentrate on the demanding tasks involved in advising customers. This is accompanied by improved customer satisfaction and higher productivity.
The volume of time and monetary savings scales with the number of incoming tickets. In the future, the DATAGROUP Service Desk can use this intelligent solution for similar processes and roll it out for other customers and their use cases.
»Our solution is very lightweight, which allowed for easy integration.«
– Karsten Schmöker, Authorized Signatory & Director Application Engineering, Almato AG