INTRODUCTION ============ The Agentic NOC introduces an AI-native, agent-based NOC architecture that transforms network operations from reactive cost centres into autonomous, strategic enablers. It addresses one of the telecom industry’s most critical challenges: operating increasingly complex, always-on networks with legacy, reactive NOC models that can no longer keep pace. ====================================================================== As 5G, cloud-native architectures, edge computing and IoT drive unprecedented scale and dynamism, communications service providers (CSPs) face rising operational costs, prolonged outages, alarm fatigue and mounting pressure to deliver flawless service while meeting sustainability mandates. Today’s NOC still operates as a war room: thousands of multi-domain alarms, manual correlation and hours to resolve. Even modern AIOps stops at recommendation, leaving humans to drive every step. Built on TM Forum’s Open Digital Architecture (ODA) and aligned with Agentic AI, Intent-Driven Autonomous Networks and Autonomous AI Control Loop (AACL) principles, the Agentic NOC deploys a collaborative “agent-of-agents” ecosystem of multi-domain intelligent agents spanning network, service, field force, security, analytics and virtual operations. Key capabilities include autonomous fault prediction and closed-loop remediation, intelligent incident management, AI-driven field-force optimisation, zero-trust security enforcement and a power-saving advisor that reduces energy consumption without compromising performance. Together they enable CSPs to achieve Autonomous Networks Level 4 and beyond; self-healing, self-optimising and sustainable network operations in which the network senses, reasons and acts inside human-defined guardrails, restoring service across RAN, transport and core in under 60 minutes with no ticket, bridge call or customer impact. Standards Foundation Rather than bolting AI onto a single vendor’s tool, Catalyst 924 is anchored in TM Forum assets so the blueprint is portable and reusable across CSPs. The architecture conforms to the Open Digital Architecture (ODA), the Autonomous Networks Architecture, the Autonomous AI Control Loop (AACL) and the IG1220 closed-loop models, and draws on the intent and closed-loop guidance in IG1218, IG1339 and IG1501. It is realised through TM Forum Open APIs, including TMF921 (Intent Management), TMF915 (Customer Management), TMF642 (Alarm Management), TMF701 (Process Flow), TMF702 (Resource Activation), TMF656 (Service Problem Management) and TMF913, providing a standards-conformant, verifiable path against Autonomous Networks Level 4 criteria. Ecosystem and Proof Points The Catalyst is demonstrated through two integrated use cases on real operator data: (1) intelligent fault correlation across RAN and transport, and (2) transport/IP anomaly detection for mobile and fixed-wireless networks, surfaced through a unified operator dashboard that links technical events to business outcomes in a single view. Champions T-Mobile USA, Türknet, Axian, Telefónica and Digital Nasional Berhad contribute production network and OSS data across North America, EMEA and Asia-Pacific, while participants Nagarro, Enfec, Snowflake, DigitalRoute and Sutherland deliver the platform and agent stack, specialised agents running over a shared Snowflake data foundation, mediated by DigitalRoute, with explainability and human-in-the-loop guardrails throughout. Quantified Outcomes Validated against real operator data rather than synthetic benchmarks, the solution targets up to a 60% reduction in mean time to repair (MTTR) and mean time to detect (MTTD), approximately 70% faster fault identification, around 40% fewer incidents escalated to human operators, and 15–20% RAN energy savings, the single largest Scope 2 lever available to operators. At a single Tier-1 operator these translate into multi-million-euro annual opex savings, materially better SLA conformance and improved CSAT; replicated across the global Tier-1 base, even partial adoption represents an industry prize measured in billions of euros and gigawatt-hours of energy avoided each year. The result is a production-ready blueprint for ANL4+ that participating operators and vendors can carry forward beyond the show.