If you have ever been on a major telecom incident bridge, you know the feeling. Alerts everywhere. Dashboards disagreeing. Smart people doing manual stitching at 2am.

Automation started as relief for repetitive work. What is happening now is more human than technical: operators are trying to protect time, reduce burnout, and make networks more resilient by letting systems handle more of the routine thinking, not just the routine clicking.

Industry bodies describe the destination as zero-touch and autonomous networks, where end-to-end management becomes increasingly automated across the service lifecycle. [1]

The evolution in 3 steps

1) Task automation

Runbooks, scripts, workflows, ticket routing, provisioning templates. Useful, but often brittle and hard to scale across domains.

2) Assisted operations

AI helps humans detect patterns, prioritize incidents, suggest probable causes, and recommend next actions. Humans still drive decisions.

3) Closed-loop automation

Systems detect, decide, and act within guardrails (policy, approvals for high-impact actions, canary rollout, rollback). This is where operators start seeing step-change improvements.

ETSI’s ZSM reference architecture is one of the clearest blueprints for building these loops in a service-centric, end-to-end way. [1]

Why this is accelerating now

5G made automation a core requirement, not a “nice to have”

3GPP has specified a network automation architecture to leverage 5G information exposure and network data analytics, aiming to improve performance, assurance, and resource utilization, and enable self-optimization and automated management. [2]

Intent is becoming the control language

Instead of hard-coding “recipes,” intent-based automation focuses on outcomes and constraints, letting the system generate solutions inside guardrails. TM Forum describes intent as enabling a level of autonomy where the system can adapt behavior and generate solutions rather than just follow human-defined recipes. [3]

Autonomy is becoming measurable

Operators are moving from “we are automating” to “we can prove our autonomy level.” TM Forum launched ANLAV to let CSPs independently validate autonomous network levels, supporting progress toward Level 4. [4]

What this looks like in the real world (by region)

Europe: autonomy becomes a program with milestones

Telefónica is publicly positioning autonomy as a structured journey and has shared that it has two Level 4 use cases in Brazil and Germany as part of its Autonomous Network Journey program, aligned with TM Forum’s ANLAV direction. [5]

Vodafone has also publicly linked Open RAN programmability with automation, describing efforts toward “one-click” customer connectivity using Open RAN automation concepts. [6]

Denmark has provided one of the clearest public proof points. Ericsson states that TM Forum verified Level 4 autonomy for a high-value scenario focused on automating RAN energy efficiency optimization, with the solution in production on TDC NET’s network since May 2024. [7]

North America: AI moves closer to live operations

Verizon has publicly described infusing AI into the network and accelerating Open RAN innovation, including the role of the RIC and rApps to improve network outcomes such as coverage, capacity, efficiency, and service quality. [8]

More recent reporting also highlights Verizon using AI inside its commercial network to manage power use and performance, reflecting how AI is moving from pilots into operational control loops. [9]

Middle East: architecture foundations plus “agentic” direction

In the Middle East, the story often starts with clean architecture foundations. TM Forum announced that stc Group became the first carrier in the region to achieve “Running on ODA” accreditation, signaling progress toward modern, composable architectures that make automation easier to scale. [10]

In parallel, a joint e& and AWS whitepaper discusses “agentic AI” and its relevance to multi-step workflows and decision-making, a concept that maps naturally to automation loops in operations. [11]

Africa: pragmatic, ROI-driven automation

Safaricom has publicly stated that, as part of its digital transformation program, it has developed AI/ML big data and analytics capabilities using AWS, providing the data foundation needed for dependable automation. [12]

MTN’s public reporting on autonomous network trials highlighted practical closed-loop style use cases including “sleeping cells self-healing” on South Africa live networks, a classic bounded automation loop with clear operational impact. [13]

A simple way to explain telecom automation without the buzzwords

Most automation that delivers real value follows the same loop:

• Sense: telemetry, logs, KPIs

• Reason: analytics, correlation, prediction, intent and policy 

• Act: orchestration, configuration changes, scaling, healing actions

• Learn: continuous improvement, updated policies, model refinement 

This “sense, reason, act, learn” thinking is exactly what ETSI ZSM was designed to support through loosely coupled management services that can be composed end-to-end. [1]

What is still hard (and why that is normal)

Data truth

If inventory, topology, and service models are inconsistent, automation becomes risky fast.

Safety and blast radius

Closed-loop is only valuable when it is trustworthy. That means approvals for high-impact actions, controlled execution, and fast rollback.

Cross-domain reality

Most real incidents cross RAN, transport, core, cloud, and OSS layers. End-to-end frameworks exist because operations is end-to-end. [1]

Practical takeaway for CSP leaders

The winners will not be the operators who “automate the most.” They will be the ones who automate the right loops first, with clean data foundations, clear intent, and safe guardrails.

If you are building a roadmap, start with the use cases that teams actually feel:

• fewer repeat incidents

• fewer manual changes 

• fewer late-night escalations

• faster recovery, with less heroics

REFERENCES

[1] ETSI GS ZSM 002 V1.1.1, Zero-touch network and Service Management (ZSM); Reference Architecture. https://www.etsi.org/deliver/etsi_gs/ZSM/001_099/002/01.01.01_60/gs_ZSM002v010101p.pdf

[2] 3GPP, Network Automation Enablers in 5GS (CT3). https://www.3gpp.org/technologies/nae-5gs-ct3

[3] TM Forum, Intent-based Automation. https://www.tmforum.org/learn/topics/intent-based-automation/

[4] TM Forum Inform, DTW Ignite: New AN Level assessment service validates progress toward Level 4. https://inform.tmforum.org/features-and-opinion/dtw-ignite-new-an-level-assessment-service-validates-progress-toward-level-4

[5] TM Forum Inform, DTW Ignite: Telefónica shares Level 4 AN progress. https://inform.tmforum.org/features-and-opinion/dtw-ignite-telef%C3%B3nica-shares-level-4-an-progress

[6] Vodafone, Vodafone working on Open RAN automation to provide customer connectivity in just one-click. https://www.vodafone.com/news/newsroom/technology/open-ran-automation-provide-customers-connectivity-in-one-click

[7] Ericsson, TDC NET and Ericsson achieves industry first certification from TM Forum of Level 4 autonomy. https://www.ericsson.com/en/news/2025/6/tdc-net-and-ericsson-achieves-industry-first-certification-from-tm-forum-of-level-4-autonomy

[8] Verizon, Verizon infuses AI in the network, accelerates Open RAN innovation. https://www.verizon.com/about/news/verizon-open-ran-innovation

[9] TelecomsTechNews, Verizon integrates AI into network operations and edge infrastructure. https://www.telecomstechnews.com/news/verizon-integrates-ai-into-network-operations-and-edge-infrastructure/

[10] TM Forum, stc Group achieves Running on ODA accreditation. https://www.tmforum.org/news-insight/newsroom/stc-group-achieves-running-on-oda-accreditation/

[11] e& and AWS, The Rise of Agentic AI in the GCC (whitepaper PDF). https://www.etisalat.ae/content/dam/eand/assets/img/general/e-aws-whitepaper-final.pdf

[12] Safaricom, Safaricom set to transform Kenyan consumers’ digital experience on AWS. https://www.safaricom.co.ke/media-center-landing/press-releases/safaricom-set-to-transform-kenyan-consumers-digital-experience-on-aws

[13] Light Reading, Leading Innovation in Africa: MTN completed key autonomous networks trials with Huawei. https://www.lightreading.com/business-management/leading-innovation-in-africa-mtn-completed-key-autonomous-networks-trials-with-huawei