Autonomous Networks Level 4: From Concept to Reality

Published in Telesemana

by Pablo Bertrand

Published in Telesemana by Pablo Bertrand

Just a few years ago, the concept of autonomous networks was still regarded in the industry as an idealistic topic with little relevance for productive operators focused on different priorities. It is over the past year that it has gained traction, to the point of becoming a central theme at an event of the scale of DTW Ignite 2025. And this comes as no surprise: at a time when artificial intelligence has advanced enough to begin materializing the vision of autonomous networks, a market such as telecom cannot ignore a proposition that offers relief to increasingly tight profit margins.

TM Forum classifies network autonomy into six levels: from fully manual control (level 0) to complete autonomy with no human intervention whatsoever (level 5). A 2024 report placed the average maturity of operators across four continents at around 2.5, which in practice means having specific automations in place (in the form, for example, of OSS modules) but with a need for constant supervision. In contrast, level 4 autonomy requires the network to self-manage, self-optimize, and self-heal, translating business objectives into technical actions, reducing operational costs by more than 50 percent, while improving user satisfaction and optimizing energy consumption.

From a technical standpoint, the leap to level 4 requires integrating contextual intelligence capabilities and modular architectures aligned with the Open Digital Architecture. However, the greatest challenge appears to lie from the executive perspective, which involves aligning governance, investing in talent, and establishing clear KPIs that measure the return on automation.

Toward Active Management

The TM Forum Blueprint Industry Level 4 describes level 4 as a "complex, multi-domain environment in which the system enables decision-making based on predictive analysis or active closed-loop management of service- and customer-experience-oriented networks, through AI modeling and continuous learning." To achieve this, automation must limit manual interventions to physical hardware replacement or repair. This model relies on distributed digital twins and AI models, orchestrated through an autonomous networks map and evaluated by the AN Level Assessment Service, which objectively compares operator maturity.

At the end of last year, TM Forum proposed a roadmap spanning from 2025 to 2027 for the implementation of 20 high-value scenarios (such as fault management in core and RAN, zero-touch provisioning, and dynamic traffic optimization) and extends the vision to 2030 to cover more complex end-to-end use cases. This roadmap is structured across three hierarchical layers: business (translation of service intents), service (multi-domain orchestration), and resource (autonomous components), each supported by key business, service, and capital indicators. In this way, operators have a clear framework for planning investments, designing pilots, and measuring results in both financial and customer experience terms.

In Latin America, the roadmap toward level 4 starts from a mid-level maturity point, where the heterogeneity of legacy infrastructures and the dispersion of data across proprietary silos define the main bottlenecks. Nonetheless, local initiatives are moving beyond traditional pilots. Telecom Argentina, for example, created an Autonomous Network Transformation Office that launched high-impact use cases in residential Wi-Fi, successfully closing measurement and action loops that reduced complaints related to failures and accidental resets. Telefonica, through its Autonomous Network Journey program, participates in the evaluation of core and IP fault management, laying the groundwork for extending autonomy to the RAN. Antel Uruguay, for its part, participated in the first phase of these pilot tests, demonstrating an early commitment to network automation.

Cases to Follow

What occurred at DTW Ignite 2025 reinforced this trajectory. In June, TM Forum presented the level 4 implementation guide and the Autonomous Networks Level Assessment and Validation service, while China Mobile demonstrated its capability for real-time automatic fault detection and remediation, and Telefonica showcased its Fractal project for intelligent 5G planning. These demonstrations confirm that level 4 is more than that earlier idealistic concept and that operators should already be developing their roadmap to advance along this path alongside the rest of the industry.

The convergence of clear definitions, validated use cases, and evaluation tools presents a scenario in which the region could move from partial autonomy to levels 3 and 4 in little more than one technology cycle. However, a balance is needed between technical ambition and local governance, with the understanding that each country and each operator faces different regulations, budgets, and connectivity demands. This challenge is as broad as it is difficult to address, given that it entails a complete paradigm shift in the telecom business: to this day, technology decisions are made by different teams according to their own budgets and objectives, generating technology silos that do not communicate with each other and only perform the basic integrations required. A clear example can be seen in the parallel inventories that operators maintain by technology, instead of a single system capable of operating across all domains. The historical progression leaves us with technologies layered one on top of another, posing an enormous challenge when attempting to modernize the network while continuing to operate it.

The counterproposal is that, instead of designing systems from the bottom up, they should be proposed from the top down, with strategies and operational policies aligned across the entire organization. Examples of this capability can be seen in greenfield operators that did have the opportunity to plan the entire network from scratch, and that today are capable of orchestrating and optimizing the network end to end, such as Rakuten Mobile in Japan. Planning, budgeting, and operations based on organizational strategies mark the path toward the high automation sought in level 4, which enables the reduction of operational costs, the decrease of incident response times, and a closer approach to service excellence for subscribers.

Tags: 5G, ANTEL, automation, Rakuten Mobile, autonomous networks, Telecom, Telefonica, TM Forum