AI-Supported Root Cause Analysis in Automotive Quality Problem Solving: A Cross Domain Literature Review

Patrick Kaftan, Carsten Lanquillon, Nicolaj C. Stache

Abstract: This paper explores the integration of AI-supported root cause analysis (RCA) across various domains, with a particular focus on automotive manufacturing. It reviews key methodologies, including natural language processing (NLP), machine learning, and knowledge-graph techniques, that enable the extraction of actionable insights from unstructured data sources such as maintenance logs and defect reports. The study examines how AI can identify fault patterns, diagnose root causes, and suggest corrective actions through a three-stage pipeline of knowledge extraction, representation, and causal inference. Despite the technical advances reported, significant gaps remain in the practical integration of these methods into established quality management systems (QMS), guided by standards such as ISO 9001 and IATF 16949, and in leveraging process audit data to enhance fault tracing and prevention. Throughout, the paper highlights challenges related to data quality, explainability, and the lack of clear frameworks for incorporating AI into formal problem-solving cycles. Future research is encouraged to address these limitations, particularly through the development of standardized terminologies, deeper integration of process audits, and improvements in explainability to meet regulatory and operational needs.

External IDs:dblp:conf/hci/KaftanLS25