Traditional Business Process Management (BPM) faces significant challenges in addressing the increasing complexity, uncertainty, and dynamic nature of modern business environments due to its reliance on static rule-based automation approaches. Although Artificial Intelligence (AI) has been increasingly integrated into BPM through Intelligent Process Automation, process mining, and predictive analytics, most existing solutions remain focused on task-level automation and have not achieved process-level autonomy. This limitation highlights the need for autonomous workflow architectures capable of adaptive decision-making and dynamic process orchestration. To address this gap, this study proposes a Multi-Agent Workflow System (MAWS) as an Agentic Artificial Intelligence-based framework for autonomous Business Process Management. The research adopts the Design Science Research (DSR) methodology, encompassing problem identification, artifact design and development, demonstration, and evaluation through five complementary validity dimensions: instrument validity, technical validity, design validity, purpose validity, and generalization. The evaluation results indicate that MAWS improves process execution efficiency, reduces human intervention requirements, and enhances processing accuracy compared with traditional Robotic Process Automation (RPA) approaches. Furthermore, the autonomous agents embedded within MAWS demonstrate capabilities for goal-oriented reasoning, multi-step planning, self-correction, and adaptive workflow orchestration under dynamic operational conditions. The primary contribution of this research lies in the development and validation of a BPM architecture based on multi-agent reasoning that bridges the gap between Agentic AI theory and enterprise-scale autonomous workflow implementation. The findings suggest that Agentic AI has the potential to transform BPM from rule-driven automation into goal-driven autonomous operations, while emphasizing the importance of robust AI governance mechanisms to ensure transparency, accountability, and operational trustworthiness.