Achieving autonomous networks: why discipline, not speed, defines the leaders

In the telecommunications industry, autonomous networks have emerged as a strategic north star. Most operators now anchor their long-term transformation plans on the TM Forum Autonomous Network (AN) framework, aiming to progress from fully manual operations (L0) to complete end‑to‑end autonomy (L5). The potential benefits are significant: quicker fault resolution, more resilient operations, lower operating costs, and networks capable of responding to demand in near real time through closed-loop automation.

Yet, despite widespread alignment on the destination, progress remains uneven. For many operators, the journey has stagnated at Levels 1 and 2. At these stages, automation is mainly limited to specific use cases such as improved monitoring, alerting, and task automation. Meanwhile, root-cause analysis and corrective actions still largely depend on human intervention. This creates a widening gap between strategic ambitions and operational realities, raising important questions about readiness, capability development, and the path to scaling autonomy across the network.

Building autonomy the hard way: discipline and trust

Autonomous networks offer substantial benefits, including a 20 percent boost in operational efficiency and an 18 percent decrease in network operating costs. Furthermore, telecommunications companies have reported significant improvements in fault detection, network scalability, and flexibility, along with a noticeable reduction in network downtime and outages [1].

The main obstacle to advancing autonomous networks is not a lack of belief, but a lack of clarity. It is about understanding what true autonomy requires and recognising where operators currently stand. Autonomous capability is not achieved through isolated initiatives; it must be deliberately developed, objectively evaluated, and actively managed to maintain operational performance at scale.

Yet, self-assessment often remains a challenge. In many cases, deploying a single tool, automating a workflow, or introducing a dashboard can lead an organisation to declare itself “autonomous.” In reality, maturity rarely develops evenly. An operator might achieve Level 4 in RAN optimisation while remaining at Level 1 in transport or service provisioning, constrained by legacy architectures and vendor fragmentation. These imbalances underscore a key truth: autonomy does not come from isolated successes, but from capabilities that operate cohesively across network domains and operational layers, guided by a structured, domain-aware maturity roadmap.

Industry data reinforces this gap between perception and reality. A 2023 report indicates that only 6 percent and 1 percent of telcos have reached Level 3 and Level 4 maturity respectively, despite broader claims of progress [2]. Before devising ambitious autonomy strategies, operators must therefore begin with an honest, rigorous assessment of their current state.

Make haste but slowly

Those who succeed in progressing toward Level 4 autonomy tend to follow a common pattern. They prioritize discipline over haste, progressing in deliberate phases rather than attempting radical leaps. Their focus is on operational readiness and governance, not on marketing narratives.

In practice, the journey occurs in three distinct phases. First, operators establish a robust data foundation by creating a unified network data layer that consolidates telemetry across domains and makes it accessible in a consistent, near-real-time manner. This is essential for any higher level of autonomy.

Second, they automate high-volume, low-risk tasks such as fault triage, configuration validation, routine remediation, backups, and compliance checks. At this stage, clear policies define where automation applies, where human oversight is necessary, and how actions can be reversed.

Third, with dependable data and proven automation, operators can introduce predictive and prescriptive capabilities. Networks begin to anticipate problems, evaluate trade-offs, and take context-aware actions, marking the move toward Level 4 autonomy.

Trust the process

Throughout this progression, trust emerges as a critical enabler. Systems that perform actions traditionally handled by humans must gain organisational confidence. For instance, Experienced NOC engineers, whose expertise has long ensured network stability through manual intervention, are understandably cautious about delegating decisions to automation without clear accountability and control. Without deliberate change management and clarity on responsibilities, even the most advanced technical systems risk failing to reach their full potential.

References

[1] [2] Capgemini – Networks with Intelligence Why and how the telecom sector should

accelerate its autonomous networks journey