Role-Based Agentic AI for Intent-Driven Network and Service Orchestration

The paper introduces a role-based multi-agent architecture (MAS) designed for end-to-end intent orchestration in telecommunications, addressing the challenges of integrating Business Support Systems (BSS) and Operations Support Systems (OSS). The architecture features a hierarchical four-layer system that includes leadership, service, and resource agents, which are dynamically instantiated to meet intent requirements. This framework enhances autonomous network management by enabling coordinated planning and service delivery, thereby facilitating scalable and accountable intent-driven orchestration in complex network environments.