Today, numerous manufacturers from various sectors operate large fleets of devices and complex production facilities that generate more and more data. With the help of artificial intelligence in data analysis, such companies can significantly optimize their processes. The Fraunhofer Institute’s project “ML4P - Machine Learning for Production” supports this with an approach to machine learning that consists of a scientific process model and software tools based on it. For example, companies can analyze machine data to discover hidden connections and optimize production processes. Thanks to their ability to learn, the tools can also continuously improve production.