This paper presents a general concept for an in-line quality control system and the basis for component pairing using the example of a pressure valve. The aim is to improve the functional quality of the product while increasing dimensional tolerances to reduce waste. The proposed system requires an end-of-line (EOL) functional test and collects pre-existing sensor data from the production line. This data is used to train machine learning models to identify correlations between measurements and EOL test results. The system uses this information to predict future EOL test results. Anomaly detection and root cause analysis is performed by comparing predicted results with actual measurements. To improve the data set, additional sensors are integrated into the identified production steps. Once parameters with a high influence on the product function have been identified, these should be used to find ideal pairs of components with favorable parameter combinations in order to improve functionality. The EOL test is then used for validation.