AI-Driven Data Governance: Intelligent Master Data Harmonization Across Heterogeneous Cloud Architectures

Abstract

Manual data governance and Master Data Management (MDM) approaches are no longer effective as enterprise data increasingly grows in volume and spreads across multiple cloud and hybrid environments. Data consistency, regulatory compliance and entity resolution are major challenges in modern organizations, especially when dealing with multiple cloud architectures. This study explores the adoption of an AI-based data governance paradigm purposefully designed for intelligent master data harmonization. The study adopts a highly scalable hybrid architecture, which combines machine learning algorithms for automated metadata discovery, probabilistic models for entity resolution, and large language models for semantic schema matching. The research compares operational performance with various cloud nodes and offers a detailed result analysis of two important parameters: entity resolution accuracy (precision, recall, and F1-score) and data harmonization latency with increasing transaction volumes. The empirical results show that the proposed framework based on artificial intelligence reaches an F1 score of 93.4% for entity resolution, which is higher than the results of the traditional deterministic rule-based systems. At the same time, it reduces latency for data cleansing in high-volume multi-cloud deployments, performing complex deduplication pipelines in just a few seconds instead of hours or days, compared with legacy architectures. The results demonstrate how AI-driven MDM pipelines can shift from being a reactive approach for governance to a proactive, scalable, and independent ecosystem that delivers data integrity and robust regulatory compliance in complex enterprise environments.