The fifth and final application example is called 'predictive maintenance'. For this, a system collects various data, including process data and machine data, but also external data, such as weather or traffic data. This real-time information is evaluated using big data technologies and correlated with empirical values from previous functional failures. This allows the system to identify patterns of when failures occur. It is constantly learning and can predict and thus prevent functional failures better and better. This application scenario is already a reality in 27% of all companies. The majority of those surveyed are planning or discussing its use.