Artificial intelligence for enhanced master data quality management in ERP systems

Abstract

This research project investigates Artificial Intelligence (AI) capabilities that can be applied to Master Data Management (MDM) systems within Enterprise Resource Planning (ERP) environments to enhance master data quality. MDM systems are pivotal in managing high-quality, reliable master data within organisations, particularly in ERP systems where data integrity directly impacts business processes and decision-making. Recent studies underscore the growing interest in automating MDM tasks using AI. However, challenges persist, such as managing master data from diverse sources, which often results in duplicates and inconsistencies and leads to bad business decisions. This research project addresses the importance of maintaining high-quality master data by exploring the potential of AI to address data-related challenges effectively. As AI technology advances, leveraging its capabilities holds promise for streamlining MDM processes and optimising business operations in ERP environments. The primary research artefact is an AI-Enabled MDM Framework that categorises MDM functions to appropriate AI data management capabilities and provides suggested AI tools and methods identified from the literature. The framework was developed using a Systematic Literature Review (SLR) approach with a Design Science Research (DSR) methodology. It categorises MDM functions and tasks to appropriate AI data management capabilities. The research artefact was created by evaluating and combining the findings from both Systematic Literature Reviews. To evaluate the findings from the Systematic Literature Reviews, a survey with corresponding questions related to the research topic was presented to data management professionals and analysed. The most prominent MDM functions were defined by classifying the MDM functions identified from the Systematic Literature Reviews according to the highest number of references in the literature and the highest percentages in the survey responses. The most prominent AI capabilities were also defined by classifying the AI data management capabilities according to the highest percentages in the survey responses and the highest number of references in the literature. The usefulness of the research artefact as a DSR contribution is demonstrated by presenting case studies of how the framework can be applied to Master Data Management challenges. The findings can be applied to future research as a theoretical framework for research in AI and MDM, and industry professionals can also use it as a reference framework for integrating AI tools into MDM systems