Understanding defect causes is essential to improving quality. In modern manufacturing, this requires moving beyond correlations to identify true cause-effect relationships within extensive production data. This session will demonstrate how Siemens applied Causal AI to PCB/A manufacturing, revealing the contributions of design and process parameters to defects detected in AOI. With a holistic, object-centric view of the production line, hidden dependencies across process steps became instantly visible. We will share key achievements, pitfalls, and prerequisites and explain how Siemens cut error rates below the stringent 5-Sigma level, paving the way for autonomous, data-driven quality.
