Applying Artificial Intelligence in the Smart Factory: Lessons Learned from real-world use cases

The smart factory is a key concept of Industry 4.0, in which the manufacturing system is fully connected. This connection results in a large amount of data being generated. Artificial intelligence (AI) aims to create machines that are as intelligent as humans. Nowadays, machine learning is the most sophisticated approach within AI, essentially describing the possibility to find patterns in large amounts of data to, for example, predict machine failure. Thus, the potentials of AI in smart factories are plentiful and are frequently reported on in literature. However, organizations who want to deploy AI in their real-world smart factories face multiple challenges such as identifying the relevant data, missing management support, or headwinds from the workforce. In this paper, we report on the lessons learned from applying AI in the smart factory in over 50 real-world use cases. We conduct a research world café and 13 interviews to consolidate eleven lessons learned. We structure these lessons learned based on the common conception of people, technology, and organization. Our findings allow the research community to reflect on possible future research directions, and allow practitioners to avoid pitfalls when conducting AI projects in smart factories.