USU has put it to the test: Artificial intelligence accelerates service management processes

Contents
USU Software AG has conducted a test to find out how much artificial intelligence affects service management. In short, the result can be described as "positive". In the test, USU took a close look at resource consumption, efficiency, accuracy and the differences between the results of classic machine learning and deep learning.
USU, ML and DL
Machine learning (ML) and deep learning (DL) in service management - and everywhere else - build on each other. Deep learning is a form of machine learning that is the best possible imitation of the structures of the human brain. In comparison, machine learning is basically like an AI putting together a puzzle, while deep learning allows the AI to build the puzzle itself.
The implementation of deep learning offers new possibilities for service management, but it can also lead to higher consumption of resources and capital. USU conducted a test to find out how the switch from machine learning to deep learning in service management works.
A data set with 100,000 tickets served as the basis.
The differences start with the pre-processing of the tickets. Deep learning reads the relevant columns, while pre-processing in machine learning depends on the composition of the tickets. This is where the first savings in time and effort are made.
However, this applies more in the long term. Training in deep learning takes longer. However, the models can be used again and again, whereas in machine learning, new models have to be created again and again.
Optimizations by USU in model training
Machine learning has its limits. A model is only as good as the tasks it is supposed to perform. In the field of service management, new tickets are constantly being created. These tickets require new models. In machine learning, a new model replaces the old one and overwrites the previous knowledge.
There is no consecutive training, so to speak. What the AI has learned before is discarded and relearned with additional, new learnings. This is why the AI regularly starts from scratch.
Deep learning, on the other hand, retains the knowledge already learned and expands it with new input. DL inserts new tickets into the existing model, thereby expanding the model and its own knowledge.
This optimizes the entire process. Instead of regular training sessions lasting several hours, short-term enhancements take place. As USU reports, the time saved by deep learning compared to machine learning in service management is more than 130 hours.
At the same time, the test showed that the accuracy of the model is higher in deep learning. The result is an increase in efficiency with less time required.
Learn service management with skill it
Artificial intelligence is an enrichment for service management. You will also discover this in our seminars, which show you how to create Learners organizational structure. Take a look at our three-day ITIL® 4 Direct, Plan and Improve course, which teaches you how to integrate continuous improvement into a system. Or take part in our ITIL® 4 Drive Stakeholder Value seminar to achieve maximum value.