From laboratory to real life: Fraport’s approach to applying artificial intelligence in airside operations and ground handling

Artificial intelligence (AI) has been gradually finding its way into various areas of our life in recent years including air traffic and airport management. Establishing a basic understanding of the characteristics of this specific class of algorithms and the associated conceptual differences to classic information technology (IT) solutions is increasingly proving to be a critical success factor when introducing and running AI solutions in a safety- and security-critical environment such as commercial aviation. This paper explains how Fraport has been establishing a multilevel organisational AI approach. The approach is demonstrated at the example of the project on establishing a more precise prediction of arrival times at the gate (Estimated In-Block Time [EIBT]) which was used as a catalyst in this process. Fraport’s approach stretches across the entire solution development process, from understanding the problem and determining the room for improvement in an application-oriented data lab format called the Corporate Analytics Centre (CAC), all the way to developing a full-scale IT solution for operational use in the ‘IT Factory’. The practical experience of the first project has shown that not only technical challenges have to be solved during the development and implementation of the first AI solutions but it was also clearly a matter of establishing trust in AI solutions on various hierarchical levels on the user side.