Automated ticket routing using artifical intelligence

The management and handling of tickets is one of the core functions many of our customers face. Typically, internal and in many cases external staff are provided with an easy to use user interface in speedikon C, that allows for the creation of various types of tickets. Given the notion of “mobile-first” there are also Apps available for ticket creation.

In order for these numerous incoming tickets to be handled effectively and efficiently by the correct person in charge, it is vital that the tickets have been categorized correctly. Nowadays many customers leave this vital categorization to the person creating the ticket. This means additional work for the ticket creator. Even worse, in many cases staff are unsure about the correct category and therefore select the wrong one – leading to a costly manual post processing of tickets.

In order to solve this problem, our development staff have teamed up with the colleagues from our Future.Lab. Using artifical intelligence and machine learning, they have created a tool that automatically analyzes and categorizes tickets. The trained system has a profound understanding of the German language (other languages are also possilble) and the meaning of words. This allows for the AI to understand even complex wordings as well as unstructured tickets. The AI-tool analyzes each and every incoming ticket in terms of it’s content and the problem at hand. The tickets are then automatically allocated into the different categories, defined by the customer. This further alllows for each ticket to automatically be assigned to the correct person or team. In case something goes wrong and a ticket is assigned incorrectly, the user can always intervene and change the category. In this case, the AI will pick up on that and use this information to improve it’s algorithms. Especially when it comes to very customer-specific settings, the AI can further be improved by using additional data.

Based on this approach, we are currently working on even more advanced features. One of the current research topics within the Future.Lab is automating ticket responses. The idea is that if the AI recognizes a ticket to have been processes before, it can then provide the user with the correct information and if needs be trigger the according internal processes.