AI Orchestration of Multi-Agent Systems for Enterprise Operations

This thesis examines the orchestration of multi-agent systems (MAS) within enterprise operational environments, with particular emphasis on the automation of procurement processes within enterprise resource planning (ERP) systems. Leveraging AI-driven orchestration capabilities enabled by Copilot Studio, the research proposes a framework in which autonomous agents collaboratively coordinate, execute, and optimize purchase requisition workflows. The study analyzes agent coordination strategies, task allocation models, and decision-making mechanisms aimed at improving operational efficiency, reducing manual intervention, and ensuring compliance with organizational policies. Furthermore, the integration of MAS within ERP ecosystems is investigated to identify key architectural design considerations, potential bottlenecks, and opportunities for intelligent process automation. The proposed approach establishes a conceptual foundation for scaling multi-agent orchestration in complex enterprise contexts, offering insights into system architecture, agent interaction models, and the practical implementation of AI-driven operational workflows. This work contributes to intelligent enterprise automation research by connecting multi-agent system principles with practical enterprise implementations.