A big data-driven system for smart continuous improvement

Increasing digitalization and automation have elevated the importance of data analysis in reducing waste and costs, optimizing resources, and enhancing sustainability. However, as processes approach their optimization limits, traditional analytics become insufficient. This study introduces ImproveXpert4.0, a big data-driven platform designed to uncover actionable insights from process data through six algorithmic modules. These modules employ Machine Learning, Explainable AI, Operational Research, and Statistical methods for root cause analysis, bottleneck detection, simulation, prediction, human performance assessment, and variation prediction. ImproveXpert4.0 was developed using Design Science Research Methodology, and its algorithms were validated in real-world Manufacturing and Logistics settings. Its conceptual foundation is presented through two key artifacts: (1) the DMAIC-PDCA Structural Flowchart, which outlines its operating logic and algorithmic integration, and (2) a High-Level Architecture that details the technological infrastructure. Insights from a Narrative Literature Review and semi-structured interviews with Kaizen experts from award-winning organizations helped define the practical requirements for the system’s design. The platform contributes to both theory and practice by promoting a structured, data-driven approach to Continuous Improvement in digital organizations.