Autonomous networks mark the dawn of a new era in telecommunications
For years, telecommunications networks have been extraordinarily sophisticated systems, but they have operated in an essentially reactive manner. They were designed with foresight, deployed, and, when something went wrong, action was taken.
Today, that model is beginning to be left behind.
Growing technological complexity, the pace of digital business and the need for efficiency are driving a profound change: the shift towards autonomous networks.
An autonomous network is not simply an automated network. It is a network capable of anticipating, deciding and acting autonomously, continuously evolving to adapt to business priorities and customer needs.
Undoubtedly, one of the key enablers of this evolution is artificial intelligence, in its various forms, including traditional AI, generative AI and agent-based AI.
At Telefónica, the evolution towards autonomous networks is structured around a key strategic programme: the Autonomous Network Journey (ANJ), which since 2021 has been driving the end-to-end transformation of networks through advanced automation and the application of various AI techniques.
The programme’s ambition is to achieve advanced levels of autonomy: this means reaching Level 3.75 by 2028 and Level 4 by 2030, according to the TM Forum framework.
To achieve this, it is essential to integrate AI into all processes and align the network’s evolution with business priorities in real time. Because, ultimately, the network is not just infrastructure: it is the element that enables the best digital customer experience and facilitates new business models.
Currently, the company has more than 500 AI use cases in production, 15 of which are already at Level 4 of autonomy, thus realising a shift towards networks capable of self-configuration, self-optimisation and self-healing.
Network autonomy is no longer an option, but a competitive necessity
The environment facing operators has changed radically.
Networks combine multiple domains (radio, core, transport, IT), cloud technologies, multi-vendor environments and increasingly rapid innovation cycles.
Added to this is a new generation of services, such as network slicing and critical applications, which demand dynamic configurations and virtually immediate response times.
In this scenario, managing the network with manual or semi-automatic processes simply does not scale.
At the same time, the pressure to improve efficiency, optimise energy consumption and allocate investment more effectively is driving the adoption of smarter models.
Level 4 Autonomy addresses all of this with a clear proposition: to manage complexity through intelligence, not through increased operational effort.
From the instruction-based model to intent-driven networks
In a traditional network, devices define how to do everything: commands, configurations and step-by-step processes.
In a Level 4 network, this is replaced by defining intent.
Teams move on to setting objectives, such as improving the customer experience, reducing latency, optimising resource usage or increasing resilience.
And the system decides how to achieve them.
This is what we know as the intent-based model, which allows the network to act autonomously, evaluating options, making decisions and executing them without direct intervention.
Continuous planning enables needs to be anticipated and resources optimised in real time
Traditionally, network planning involved lengthy cycles, manual analysis and decisions which, in many cases, were already out of date by the time they were implemented.
In an L4 network, planning becomes a continuous process. The network analyses demand patterns, user behaviour, resource usage and even external factors in real time, and dynamically adjusts its decisions.
This process is supported by technologies such as Digital Twins, which enable the entire network to be simulated in a virtual environment.
Before implementing any changes, thousands of scenarios can be evaluated and the best option chosen without risk.
This approach is already a reality in various operations. In Brazil, artificial intelligence automatically generates network designs in the IP domain, drastically reducing both time and errors. In Spain, the planning of fibre deployments uses AI to optimise routes and costs. And in Germany, three-dimensional digital twins enable the network to be visualised and simulated with a level of detail that includes urban infrastructure, links and real-time traffic.
The result is clear: planning shifts from being slow and reactive to becoming a fast, precise and continuously adaptive process.
Autonomous deployment enables innovation without affecting the customer experience
The deployment of new capabilities has historically been a complex process, involving multiple teams, maintenance windows and associated risks.
In an L4 network, this paradigm shifts radically.
Deployment begins with an intention – for example, to deploy a new software version – and the system takes care of everything: it plans, executes and validates.
Before any changes are made, potential impacts are simulated. During execution, the network is monitored in real time and, if any anomalies are detected, automatic rollback mechanisms are applied.
This enables continuous deployments, with no visible interruptions for the customer.
In fact, there are already scenarios where 5G core updates are carried out without affecting the service and where autonomous workflows have significantly reduced deployment times.
From incident resolution to intelligent problem anticipation
In the traditional model, incident management is a reactive process: multiple alarms, coordination between teams and manual resolution.
At Level 4, this model is completely transformed.
The network analyses millions of metrics in real time, detects anomalous patterns and anticipates failures before they occur.
When an incident occurs, the system has already correlated the information, identified the root cause and triggered the necessary actions.
This is achieved through closed loops, where the entire cycle—detection, analysis and action—is executed automatically.
A clear example is a fibre cut. In a traditional environment, this would trigger a flood of alarms and a manually coordinated response.
In an L4 network, the system identifies the problem, assesses its impact, reroutes traffic and mobilises resources automatically, whilst operators monitor the process.
The result is a more controlled, efficient and predictable environment.
Artificial intelligence enables a more personalised and proactive customer experience
Beyond simply operating and resolving incidents, an autonomous network constantly improves its own performance.
Artificial intelligence enables the dynamic adjustment of parameters such as bandwidth, traffic prioritisation and slice configuration.
Furthermore, the network can anticipate congestion and take action before it affects the user.
This enables a shift towards a more personalised experience, where the network adapts its behaviour based on context, usage and customer needs.
Telefónica is already turning the vision of autonomous networks into reality
This model is not merely theoretical.
Telefónica is already taking firm steps in this direction through the ANJ, establishing itself as one of the global leaders in the evolution towards autonomous networks.
In addition to the use cases already deployed within the ANJ framework, it has recently been recognised by TM Forum with the “Catalyst Innovator: Voyager 2026” award for its impact, leadership and collaboration within the global innovation ecosystem through the Catalyst programme.
Smarter network operations
- Transformation of network operations centres (NOCs) into AI-native environments.
- Automation based on closed loops (detection → decision → action).
- Networks capable of self-repair using AI and digital twins.
Intelligent service orchestration and design
- Agent platforms to manage complex networks in a scalable manner.
- Intent-based orchestration.
- Automatic optimisation of network planning and investment.
Enhanced customer experience
- Use of AI to identify new drivers of perceived quality.
- Predictive models that anticipate problems before they affect the user.
- Integration of network and business data to improve the NPS.
Security and trust
- Implementation of Zero Trust architectures for agents.
- Systems that enable AI to run in secure, vendor-independent environments.
More efficient field operations
- AI-powered assistants for technicians.
- Real-time support during network interventions.
Scaling autonomy will be the major challenge of the coming decade
Level 4 networks represent a profound change in the industry.
Not only do they enable the management of growing complexity, but they transform it into a competitive advantage.
Thanks to autonomy, networks can operate more efficiently, adapt rapidly to new demands and enable innovative business models.
Ultimately, they enable a shift from reacting to anticipating, from executing to deciding, and from managing complexity to driving growth.
The challenge is no longer to demonstrate that autonomy is possible, but to scale it effectively.
This involves:
- Integrating intelligence natively, not as an additional layer.
- Strengthening key capabilities such as data, orchestration and assurance.
- Ensuring that autonomy translates into real value for the business and for the customer.
To this end, open innovation and the role of the ecosystem are fundamental.
Our collaboration with TM Forum reflects Telefónica’s active role in the global ecosystem, helping to define common standards, develop shared solutions and accelerate the sector’s transformation.
And, of course, alongside this technological transformation, the role of human talent is essential.
Engineers are no longer merely performers of manual tasks; they have become the ones who define the intent, set the rules and oversee the system’s behaviour.
Artificial intelligence does not replace human knowledge; it amplifies it.
The combination of the two is what will enable us to achieve new levels of efficiency and quality.
Telefónica and the future of smart networks
The evolution towards autonomous networks is fully in line with Telefónica’s ambition to “become the best gateway for citizens to access digital technologies”. The application of artificial intelligence, advanced automation and intention-based operating models enables the construction of networks that are more efficient, resilient and equipped to respond to the needs of customers, businesses and society.
Furthermore, this transformation contributes directly to strategic objectives such as “having the best network for accessing the most innovative technology”, “offering more and better services to customers” and “retaining and attracting the best professionals suited to the sector’s needs”, thereby reinforcing Telefónica’s technological and competitive leadership in the new digital economy.
