A scalable framework for predictive quality implementation in discrete manufacturing

The transition toward sustainable manufacturing requires a shift from a reactive approach to quality management towards predictive and prescriptive ones in order to prevent directly material and energy waste at source. Predictive quality systems, enabled by machine learning and real-time process monitoring, offer a promising opportunity to prevent defects generation, however, existing approaches are typically designed for highly digitalized environments and provide limited guidance for manufacturers operating at different levels of digital maturity, particularly small and medium-sized enterprises. To address this issue, this paper proposes a scalable framework to support the identification of the most suitable path towards quality management improvement for each organization. The proposed framework is structured around two complementary components: an assessment matrix that identifies the current digital and organizational maturity of manufacturing systems and an implementation matrix that links each maturity level to feasible analytical solutions, describing data and infrastructure requirements, decision authority and expected sustainability implications. Four progressive levels are defined, reflecting the transition from reactive quality control to real-time, closed-loop process optimization. The framework has been developed through an iterative design process combining insights from the literature on predictive quality and Industry 4.0 maturity models and its applicability is illustrated through four injection molding application cases (two implemented directly and two derived from scientific literature to represent additional maturity configurations), showing how different industrial conditions translate into different predictive quality implementation pathways.