The use of artificial intelligence (AI) in the service management of modern service organizations

A major German equipment manufacturer recently reported a sensational increase in throughput rates and quality improvements in the area of service management through the use of artificial intelligence (AI) in the service portal and service request management and associated workflows with high automation potential. Here, the support ticket rate that had to be opened and processed by service desk personnel dropped by 35%. In addition, an increase rate of 25% was achieved by means of AI-controlled automation workflows in the service request environment. What does this mean in concrete terms for processes and services, and what maturity index can be linked to this? Let's analyze the topic a little more closely.

What can AI do and what can't it do?

Artificial intelligence (AI) is considered a key technology. But the sole use of this innovative technology does not automatically make the problems and challenges for modern service organizations or companies in the digital transformation disappear into thin air. After all, the algorithm is only as intelligent as the data it learns from and implements its logic allows. In the end, it's not the use of the technology that matters, but the process and content to drive it. Accordingly, service organizations must upstream targeted ORGA projects to harness AI technology with structured data objects and data relationships in knowledge management and service support.

What is the difference between artificial intelligence (AI) and machine learning (ML)?

Artificial intelligence primarily refers to all technologies that mimic human intelligence. Machine learning is a sub-discipline within it - because it mimics learning and applying what is learned. In ML, machines or systems learn to solve tasks on their own using large amounts of data. This orientation is even more specific in Deep Learning (DL).

What claims lead to misunderstandings when using artificial intelligence?

1. Artificial intelligence basically solves all problems directly and automatically, without having to create corresponding prerequisites.
2. The use of artificial intelligence can drastically save noticeable personnel costs, as the technologies and machines are intelligent enough to do everything on their own.

What are the requirements for the application of artificial intelligence?

The basic premise of data quality in measurement and reporting is "garbage in - garbage out". AI is not intelligent on its own, but learns by means of ML from existing data, which must be compiled, structured, and provided in advance by the service organization. If this data is of poor quality, it may be used incorrectly, interpreted incorrectly, and thus used incorrectly.

An AI system operates fundamentally on three levels (E1-E3):

E1 - What do I perceive?

E2 - What can I deduce?

E3 - How should I react?

Perception takes place via sensor technology, i.e. movement data or master data. As a result, an action is fed back. A decisive capability of AI systems is therefore to draw conclusions about the state of machines or systems in the focus of observation (here related to service management, for example service portal queries or searches for knowledge objects) on the basis of data. Thus, AI does not solve all problems by itself, but is explicitly and controlled by upstream measures of data generation, structuring and integration for application in the system-side AI integration of our service management solutions.

Anyone who, as already mentioned, wants to use AI to reduce personnel costs is taking the wrong focus here. It is more important to have your own service processes under control so that they are efficient, flexible and agile at the same time. Up to 65% of costs are hidden in service processes and workflows, which could not be made transparent until now due to a lack of or insufficient technical options. Data intelligence can also help in identifying these cost drivers. Examples include approaches such as predictive analytics (PA), predictive quality (PQ), or the cross-process visualization of weak points using process mining (PM). This is an AI-based technology that can reconstruct and evaluate business processes across the board on the basis of digital traces in IT systems.

What is the recommendation for the integrative consideration and application of AI in service management?

Once the fear of contact with artificial intelligence has faded, experience must be gained. The best approach is to proceed in two stages. The first step is to test the possibilities and use cases through experimentation and to quickly make the unavoidable mistakes by means of the experiments (fail fast). The second step is to use the experience gained to roll out the successful experiments and scenarios productively (scale fast). Partial results and findings are often more important and valuable contributions than the final project result as such.

What experience have you already had with the use of AI in your service organization?

What prerequisites do you see for using AI in a meaningful and value-creating way, e.g. in the support area or to enhance the user experience? Let's talk about it!