Quality remains a crucial factor for ensuring competitive advantage and sustainability of processes in today’s market. However, despite decades of research and efforts to improve quality management, it remains a prominent issue in practice. Traditional tools mainly focus on addressing the symptoms, with low attention placed on the true causes underlying the problem. The same criticism holds for many AI tools. In contrast, Root Cause Analysis (RCA) has proven to be an effective tool in the identification and the analysis of the causal relationships among the entities involved in a system, which is the predecessor for any problem solving. Data driven RCA of production faults and defects can be implemented in different ways. Bayesian Network (BN) represents the most promising framework for characterizing processes with stochastic uncertainty, using state transition based on conditional probability. This study reviews the literature and develops a first comprehensive and domain-independent framework to guide practitioners in the implementation of RCA using Bayesian Networks.