Building Level 4 Telco Autonomy: A live discussion

"If the loop cannot self-explain, realign when an issue happens and have guardrails to resolve that issue, then you're not L4. You're mostly automation with risk."

At MWC Barcelona 2026, Rakuten Mobile announced it had earned TM Forum validation for Level 4 autonomy in energy efficiency, delivering a 20% reduction in RAN energy consumption. In this episode of Zero-Touch Live, Geoff Hollingworth speaks with Ahmed Gamal Abdelaziz, VP of automation and transformation at Rakuten Symphony, to walk through the five operational layers that underpin a genuine L4 system and what it takes to build, govern and trust these emerging systems.

Lessons learned deploying Level 4

Geoff and Ahmed unpacked what real autonomous network design looks like behind the scenes, emphasizing:

• A unified data foundation is where every serious autonomy journey starts. Networks built on siloed data can run algorithms, but isolated processing limits what those algorithms can actually learn. Bringing data into a single, accessible layer is the prerequisite for everything that follows.
• Policy and intent management is what separates goal-driven autonomy from rule-based automation. Individual policies handle specific tasks. An intent management layer orchestrates them toward the highest-value outcome, allowing the system to balance competing objectives rather than execute in isolation.
• Decision explainability is the line between real autonomy and automation with risk. A system that acts without surfacing the data points, confidence levels and risk factors behind each decision cannot be trusted, audited or improved. Explainability is a structural requirement.
• Execution intelligence includes knowing when not to act at scale. Canary deployments, pre- and post-checks, and automatic rollback logic are how a system earns the right to expand its decision-making radius over time.

The biggest takeaway from the conversation? Level 4 represents an important milestone, and evidence that a system has been designed, governed and stress-tested to operate autonomously in high-value scenarios while continuing to learn from each decision it makes.

👉 Watch the full interview to learn more about what it takes to build toward autonomous network operations for telecom.

‍