Quality assurance, zero-defect manufacturing and advances in automation are contingent on mastery of the production processes. For this purpose both technical and organisational measures are required. As far as technology is concerned it must be assured that the defined process variables remain stable or are verified continuously or at periodic intervals. For this type of process monitoring both demands and possibilities have risen substantially in recent times, as a result in particular of developments with sensors and information or knowledge processing. In this way diagnosis and monitoring of machines and systems can be enhanced and repairs based specifically on updated and improved information. These new and above all expedient possibilities are illustrated here in the description of a knowledge-based self-corrections and diagnosis system for CNC coordinate measuring machines (CMM). The expert system for reducing irregularity-related stoppages is considered not in isolation but as an important component of production-related information processing.