The ‘Evolving to full network autonomy’ Catalyst showcases a Level 4+ autonomous network, using intent management, enhanced observability, knowledge graphs, digital twins and AIOps-driven closed loops to detect, diagnose, and resolve issues autonomously. It boosts CSP service quality, drives efficiency, and sets a new industry benchmark for scalable, reliable, intent-driven networks.
Full network autonomy through intent management, enhanced observability, knowledge graphs, and AIOps-driven closed loops
Commercial context
Accelerating maturity in autonomous networks is a strategic challenge. Around $800 million in potential value could be captured in increased revenue and operational cost reduction through full network autonomy. Yet CSPs still face immense operational challenges.
First, they must deliver increasingly complex, reliable, and adaptive services to meet ever-rising customer expectations. However, as networks expand across diverse services, devices, and geographies, manual operations and traditional monitoring fall short. Service quality suffers as a consequence, and the risk of costly downtime increases. Meanwhile, customers now prioritize a seamless, uninterrupted experience.
CSPs need to radically simplify determination over whether a service can be delivered, with contextual awareness (to a specific location or customer, based on technical and commercial constraints) in near-real time. CSPs can achieve this through AI and digital twins, adopting TMF 921 for autonomous intent service management and closed loop assurance.
The ‘Evolving to full network autonomy’ Catalyst aims to prove this by building a fully autonomous Level 4+ network. Specifically, the project showcases how AI and agentic AI can be deployed across intent handling, network troubleshooting and fault resolution to anticipate network behavior and generate context appropriate closed-loop actions. Ultimately, a self-managing, self-adapting network will allow CSPs to significantly improve service reliability and quality. At the same time, it will drastically reduce operational costs by minimizing manual oversight.
Deutsche Telekom AG’s Lead Architect OSS, Sebastian Zechlin, remarks that “achieving full network autonomy will enable DT to deliver superior quality to our customers, unlock new business opportunities, and accelerate time-to-market. By enhancing operational efficiency, autonomous networks will drive significant value across our operations. We believe intent-driven networks are a key enabler for advancing automation and realizing this vision of complete network autonomy.”
The solution
The Catalyst project combines AIOps, digital twins, and AI agents to create a powerful framework for autonomous operations. Centered around intent, awareness, analysis, decision, and execution, this framework aims to transform how networks operate. Specifically, it addresses both intent orchestration and intent assurance, ensuring that network operations are not only automated but also aligned with business goals.
To achieve this, the Catalyst uses enhanced observability, knowledge graphs, and AIOps-powered closed loops. These tools work together to enable the network to detect, diagnose, and resolve issues on its own. As a result, organizations can reach new levels of operational efficiency. At the same time, customers benefit from a more reliable and responsive service experience.
AI and genAI play a critical role across multiple dimensions. For instance, they support tasks such as intent probing, identifying issues, automating troubleshooting, and enabling closed-loop operations. Additionally, they contribute to autonomous problem resolution, making the entire system more resilient and adaptive.
Digital twins are another core component. When used in conjunction with intent probing, they significantly improve service qualification and configuration. This combination increases the accuracy of service feasibility checks and reduces the likelihood of failure. It also helps test new configurations safely before deployment, which not only accelerates the sales process but also enhances the overall customer experience.
By simulating changes in a virtual environment first, this approach boosts success rates and minimizes operational risk. As AI agents are increasingly used for decision-making and action generation, the ability to safely integrate AI into the network becomes critical. Without mechanisms to validate and de-risk AI-generated recommendations, their use in live network paths would remain constrained.
The solution functions at the service layer and spans multiple domains. Consequently, it supports the entire service lifecycle, ensuring seamless integration and performance across complex network environments. The Catalyst uses TM Forum’s TMF921 Intent Management and TMF645 Service Qualification APIs. TMF921 plays a crucial role, with the catalyst focusing on the power of TMF921 probing in conjunction with a network digital twin.
The intent probing process begins when Qvantel Flex BSS, as the intent owner, sends a TMF 921 probe to Nokia's AI-powered Digital Operations Center, which serves as the intent handler. Using advanced ML models, LLMs, RAG, and an agentic AI architecture from Microsoft, Nokia's system validates the probe against a service digital twin in the unified inventory, using AI to predict traffic patterns and detect anomalies.
The proposed service is then stress-tested against expected network conditions using a network digital twin. Following these analyses, Nokia generates detailed intent reports that enable Qvantel's BSS to provide customers with data-driven service recommendations. After successful validation (with inventory customizations handled by Infosys), the process transitions to autonomous closed-loop operations.
These loops continuously monitor and correlate network data, identify anomalies, identify root cause for potential issues, and then use AI agents to generate optimization recommendations that are tested in the digital twin environment before implementation. The assurance center verifies outcomes and seeks operational approvals, creating an ongoing cycle of automatic detection, analysis, and response that maintains service quality throughout the intent lifecycle.
Wider application and value
The Level 4+ network autonomy achieved delivers benefits in business growth, customer experience, and efficiency. The solution drives business growth by enabling CSPs to unlock new monetization opportunities through innovative network services. It uses AI-driven sales automation, combined with insights from digital twins, to help CSPs confidently sell connectivity solutions. Additionally, closed-loop assurance allows CSPs to offer reliable service level agreements to B2B clients.
As CSPs adopt more advanced network operations, they can expand their service offerings and streamline enterprise sales. This not only improves efficiency but also accelerates revenue generation. For instance, CSPs can achieve a US$144 million annual revenue uplift through faster, more profitable service innovation, according to STL Partners. Moreover, Nokia reports a 2X faster time to market for new services. CSPs can also launch slice-based services in minutes, boosting competitiveness and business agility.
The Catalyst solution enhances customer experience by enabling agile, consistent, and high-quality services with minimal failures. Powered by genAI and digital twin technologies, it performs faster and more accurate service feasibility checks using intent probing. As a result, the risk of service failures is reduced, leading to higher customer satisfaction.
In addition, the Catalyst supports highly reliable operations that improve responsiveness and strengthen user trust. For example, it delivers 99% right-first-time service through precise feasibility validation – the cost of which is reduced by 68% - while 97% of potential service impacts are identified before reaching the customer. The solution also improves operational handling. It achieves 86% automated incident resolution, assigns 97% of trouble tickets to the correct team, and reduces ticket handling time by 25%, while service onboarding time is reduced by 95%.
The Catalyst solution significantly improves operational efficiency by reducing complexity and streamlining network management. Its genAI-powered closed loops continuously monitor data, detect potential issues, and automatically test fixes using digital twin technology. Manual intervention is thereby minimized, and network reliability enhanced.
The Catalyst also helps CSPs lower operational costs and better deploy resources. This not only increases organizational agility but also supports strategic growth. By using AI-driven decision-making and agentic AI, it boosts service reliability and operator productivity. According to STL Partners, CSPs reaching Level 4+ automation can save up to $650 million annually in CAPEX and OPEX. Nokia reports a 50% reduction in OPEX for new service creation and a dramatic shift in complex issue resolution - from days to just minutes.
As CSPs strive to boost revenue through advanced technologies while managing cost and complexity, operating networks with scale, reliability, and efficiency under dynamic conditions has become essential. This Catalyst demonstrates a pragmatic approach towards autonomous networks – using intent-driven design, AI, and closed-loop automation through an open, modular approach – enabling flexible, future-ready integration into the CSP ecosystem.