Login
Register
This Catalyst shows how agentic AI, intent-based automation, digital twins and TM Forum assets can help CSPs move from reactive network operations to reliable, service-aware autonomy.
Agentic NOC: AI-native operations for the autonomous telco
As communications service providers move toward autonomous operations, one of the biggest barriers is not a lack of automation, but the challenge of making automation reliable, explainable and operationally useful at scale. The Agentic NOC: AI-native operations for the autonomous telco Catalyst project explores how an AI-native network operations center can help CSPs shift from reactive, manual processes to predictive, intent-driven operations that continuously learn from network, service and customer context.
The project was named winner of the Autonomy accelerator category in the Open Innovation Catalyst Awards 2026, recognizing its contribution to accelerating practical progress toward autonomous networks.
Network operations are becoming more complex as CSPs introduce advanced 5G services, edge computing and AI-enabled capabilities. Yet many operational environments remain fragmented across network, service management and business support systems. This makes it difficult to understand the full impact of incidents, predict failures, validate customer intent or close the loop between commercial opportunity and assured service delivery.
Agentic NOC addresses this challenge by bringing together agentic AI, intent-based automation, digital twins and a common data foundation. The aim is to create a more holistic operating model, where autonomous agents can recommend, validate and trigger actions with appropriate guardrails and human-in-the-loop oversight.
The Catalyst demonstrates how an autonomous network, combined with AI-driven BSS capabilities, can accelerate CSP B2B sales and service assurance for advanced 5G slicing. The team focuses in particular on 5G AI slicing, which combines guaranteed 5G standalone connectivity with GPU computing capacity at the edge of an AI-RAN.
The use cases span intelligent customer need identification, intent validation, rapid deal closure, best-offer upsell identification, monitoring of end-user experience and guaranteed service availability. They also include agentic lead generation, negotiation and intent probing at the edge, agentic user-level incident prediction and intent resolution, and closed-loop workflows across incidents, problems and changes.
According to the project team, the Catalyst proposes a shift from fragmented network autonomy to reliable, holistic autonomy. Rather than anchoring decision-making only in network insights, the solution connects network data with impacts on end-user experience, security, service performance and business outcomes. This gives autonomous agents the context needed to operate at scale and support real-time action.
The approach also extends autonomy beyond network operations into monetization and service management. By connecting incidents, problems and configuration changes with service-aware context, the Catalyst breaks down traditional barriers between operations and service management, helping CSPs move toward true end-to-end visibility and autonomy.
“The leap from intent to impact is no longer a theoretical goal for Telenor, it is an operational necessity in a world where manual processes have become a bottleneck,” said Afnan Ahmed, Director NW Strategy and Architecture, Telenor, and Sana Khan, Director Digital Transformation Strategy, Telenor. “By leveraging intent-based automation and digital twin technology, we are turning sophisticated models into a self-governing network that removes operational barriers and drives unprecedented cost efficiencies.”
The Catalyst uses and aligns with a broad set of TM Forum assets, including TMF921 for intent probing and reporting, TMF633 for service discovery, TMF645 for service qualification against forecast network data and digital twin testing, TMF639 for resource inventory management, TMF638 for service inventory management and TMF641 for slice attachment.
The project also uses TM Forum ODA and Open APIs to scope and design integration across components, supported by assets including TMFC050, TMFC001, TMFC002, TMFC003, TMFC015, TMFC060, TMFC007, TMFC008, TMFC009, TMFC027, TMFC036, GB1086, GB1339 and a range of guides covering autonomous networks, intent, digital twins, generative AI and AI agents in telecommunications.
The team expects the solution to deliver measurable improvements across monetization, operations, customer experience and cost efficiency. These include faster service onboarding, a drastic reduction in service failures, significant OPEX savings and operational efficiency gains, and faster time to market for new services by up to 90%.
By enabling CSPs to monetize 5G AI slices, the Catalyst also points to new revenue opportunities that combine AI computing capability with assured connectivity. The project team says the approach could help CSPs cut costs by 30%, improve customer experience by shortening resolution times and enable autonomous decisions to be made with full operational and service context.
Agentic NOC shows how CSPs can use standards-based integration, intent-based automation and agentic AI to progress toward Level 4+ autonomous operations. By combining network intelligence with service, customer and business context, the Catalyst offers a practical model for moving from isolated automation to reliable, end-to-end autonomy.
For operators, the value lies in turning AI and automation into operational outcomes: faster service delivery, fewer failures, more resilient customer experiences and new ways to monetize advanced network capabilities.
