Künstliche Intelligenz im Stammdatenmanagement: Eine empirisch-literatur-basierte Analyse von Aufgaben, Technologien und Herausforderungen im industriellen Umfeld

Abstract

The steadily growing volume of data poses significant challenges for companies. As data quantities increase, so does complexity, making structured and high-quality master data management indispensable. According to the literature, master data forms the foundation of numerous business-critical processes triggered by transactions. Errors or inconsistencies in master data can generate considerable costs, as they affect large numbers of downstream processes. At the same time, the manual maintenance of master data requires substantial resources, placing a burden on departments and limiting efficiency potentials. Against this backdrop, automated and AI-supported approaches are gaining importance in sustainably improving data quality and process reliability. Innovative technologies such as Artificial Intelligence (AI) are increasingly in focus, offering the potential to support processes of data maintenance, anomaly detection, and quality assurance. However, in practice there is still widespread uncertainty about which AI technologies are suitable, how they can be implemented, and which organizational frameworks are necessary. For the empirical part of this study, a total of 14 qualitative expert interviews were conducted between June and July 2025 with professionals from industry, trade, and consulting in the DACH region. The research design followed a semi-structured guideline closely aligned with the research questions. The study addressed three core aspects: the typical tasks and activities in master data management, the AI technologies already in use, and the main challenges and potential contributions of AI to improving data quality. The findings reveal clear patterns. Master data management is considered strategically relevant in all companies, yet often lacks clear organizational anchoring, defined responsibilities, and cross-system standards. AI is currently applied mainly in pilot projects or narrowly defined use cases, such as duplicate detection or material data classification. While the potential in automation and data quality improvement is widely recognized, it remains underutilized due to limitations of existing ERP systems, insufficient investments, and a lack of expertise. The thesis provides practical recommendations and identifies success factors for the use of AI in master data management. From a scientific perspective, it extends the literature by offering empirically grounded insights into organizational, technological and cultural framework conditions. It becomes clear that beyond technological aspects, data governance and AI governance with clearly defined responsibilities, transparent processes and appropriate frameworks are crucial to ensuring that AI can serve as a strategic lever for data quality and competitiveness in master data management. Keywords: AI Governance, Artificial Intelligence, Data Governance, Data Management, Deep Learning, Machine Learning, Master Data Management

Date of Award	2025
Original language	German (Austria)
Supervisor	Patrick Brandtner (Supervisor)