Breaking Silos with Autonomous Networks: Key Insights from the TM Forum Webinar

Cross-domain autonomous networks deliver business value only when they connect customer outcomes, AI-driven operations and open standards across RAN and Transport. That was the central message of the TM Forum Global Architecture Forum (GAF) webinar, Breaking Silos with Autonomous Networks: Unlocking Business Value through Cross-Domain Autonomy, held on 16 September 2026.

The session brought together four voices from across the ecosystem:

• Priya Saxena, CTO, AI and Cloud, Zinkworks
• Roberto González Librán, Head of Observability and Automation, Orange Spain
• Jen Hawes-Hewitt, Global Industry Market Lead, Telecom Strategic Industries, Google Cloud
• Olta Vangjeli, Program Director, Autonomous Networks, TM Forum

In February 2026, Zinkworks announced a partnership with MasOrange (now known as Orange Spain), Spain’s largest telecoms operator, to deliver a Network Agentic AI solution built on Google Cloud, and many of the lessons shared in the session come directly from that work.

The appetite for this conversation was clear: 163 people joined live, and 383 more registered to watch the recording. That is a combined audience of more than 540 telecom professionals focused on one question: how do operators turn Level 4+ autonomy from an ambition into measurable bottom-line results?

In this recap, we share the five takeaways that matter most for operator leaders planning their autonomous networks journey.

Inside the Zinkworks and Orange Spain Partnership

Zinkworks is working with Orange Spain to deliver a Network Agentic AI solution, built on Google Cloud, that accelerates the operator’s move toward autonomous operations. With more than 39 million broadband and mobile lines, Orange Spain needs automation that works at national scale.

The solution is built around three use cases:

• AI-powered Root Cause Analysis: automates the diagnosis and resolution of incidents, enabling self-healing and reducing downtime and Opex.
• Real-Time Anomaly Detection: identifies unusual patterns as they emerge, so teams can intervene before service degrades.
• Network Digital Twin and Proactive Maintenance: a virtual replica of the network, continuously updated with telemetry, configuration, alarm and traffic data, that supports simulation, predictive maintenance and autonomous operations.

Google Cloud provides the data foundation. Orange Spain CTO Miguel Santos has highlighted that making key data available on Google Cloud supports a real-time network digital twin, a cornerstone of the operator’s focus on AI automation and its Customer Experience Operation Centre (CXOC). Read press release here.

Why Cross-Domain Autonomy Matters Now

Most operators have automated individual domains. The harder problem sits between them. A degradation in the RAN often originates in the IP or optical backhaul, yet each domain is monitored by separate teams, tools and vendor controllers.

This fragmentation keeps operations reactive. Engineers spend valuable time correlating alarms across boundaries, and customers feel the impact before the root cause is found.

The webinar explored how combining anomaly detection, root cause analysis (RCA) and digital twins can break down these silos. The goal is a shift from reactive domain troubleshooting to predictive, closed-loop network autonomy, delivering lower Mean Time to Repair (MTTR), guaranteed end-to-end SLAs and measurable ROI.

1. Start with the Customer and the Business, Not the Network

Cross-domain autonomy succeeds when it begins with top-down business intent. The panel agreed that operators must translate business goals, SLAs and customer experience KPIs such as Quality of Experience (QoE) into actionable network parameters across RAN and Transport.

In practice, this means an autonomous system should understand that a premium enterprise SLA or a degraded video experience in a busy cell matters more than an isolated alarm. Intent flows down from the business layer and becomes concrete policies, thresholds and priorities in each domain.

In Zinkworks’ work with Orange Spain, this alignment was anchored in TM Forum’s Autonomous Networks Business Requirements. These best practices helped ensure that the move toward Level 4 autonomy translated directly into reduced Opex and improved customer experience metrics.

2. Agentic AI and Digital Twins Turn Data into Prioritised Action

Agentic AI and digital twins move operations from reactive troubleshooting to predictive, closed-loop execution. The key is integrating real-time data from multiple domains into a single, living model of the network.

A key lesson from the Zinkworks and Orange Spain deployment was the value of building a digital twin that connects RAN topology with Transport backhaul. With that cross-domain view in place, AI agents can:

• Detect anomalies that span domain boundaries, not just within one silo.
• Identify the probable root cause, whether it sits in the radio, IP or optical layer.
• Simulate the impact of a fix before applying it to the live network.
• Prioritise self-healing actions based on customer and business impact.

The result is a fundamental change in how operations teams spend their time. Instead of chasing alarms, engineers supervise an intelligent system that resolves issues before customers notice them. Zinkworks is applying the same agentic approach to the RAN itself, as explored in Agentic SMO: Why Your Next rApp Should Build Itself.

3. Trust is Designed in, Not Added Later

Level 4+ autonomy depends as much on trust and governance as on technology. Operators will only hand decisions to AI systems when those systems behave predictably and transparently.

The panel highlighted three building blocks of trustworthy autonomy:

• Safety boundaries that define what an AI agent may and may not change.
• Human oversight at the right decision points, with clear escalation paths.
• Fallback mechanisms that return the network to a known good state if an action fails.

Technology alone does not close the gap. Organisational inertia across legacy teams is often the bigger barrier. Working with Zinkworks, Orange Spain used TM Forum’s Digital Maturity Model (DMM) to assess its organisational and cultural readiness. A key finding was that Level 4 autonomy is not only a technical challenge. It also requires a shift in data governance and operational skills.

4. Open Standards Make Autonomy Scalable Across Vendors

TM Forum Autonomous Networks frameworks and Open APIs are essential to scale autonomy beyond a single domain or vendor. Most operators run multi-vendor networks, each domain managed by a proprietary controller. Without common standards, every integration becomes a bespoke project.

Standards provide the shared language that lets cross-domain AI work consistently. To support Orange Spain journey toward Level 4 autonomy, Zinkworks and Orange Spain drew on a focused set of TM Forum assets:

For operators, the lesson is practical. Building on these frameworks reduces integration risk, protects existing vendor investments and creates a repeatable path to scale.

5. Build a Value-Led Roadmap: Quick Wins and Long-Term Bets

The most credible path to Level 4+ autonomy balances immediate returns with strategic investment. Operators face short-term business pressures, and autonomy programmes need to show value early to keep momentum and executive support.

The panel recommended a two-track approach:

• Quick wins: Prioritise high-value, low-friction use cases that deliver an immediate reduction in MTTR. These build confidence, prove ROI and help win over sceptical teams.
• Strategic investments: In parallel, invest in the foundations for cross-domain transformation, including unified data, digital twins, standards-based integration and new operational skills.

This balance keeps the programme grounded in measurable outcomes while steadily building toward fully autonomous, intent-driven operations. For a closer look at how operators can progress along this path, read Accelerating Network Autonomy through the Digital Continuum.

Moving from Ambition to Measurable Value

The message from Zinkworks, Google Cloud, Orange Spain and TM Forum was consistent: autonomous networks deliver real business value when operators break down silos between domains, teams and vendors. Success depends on aligning autonomy with customer and business outcomes, designing trust into AI from day one, building on open standards and following a roadmap that proves value at every stage.

For operators planning their next step toward Level 4+ autonomy, the work Zinkworks is delivering with Orange Spain offers a practical blueprint.

Missed the live session? Watch the webinar recording on TM Forum to hear the full discussion.

Ready to explore cross-domain autonomy for your network? Talk to the Zinkworks team about how the approach we are taking with Orange Spain, combining agentic AI, digital twins and TM Forum standards, can accelerate your journey to measurable MTTR reduction and end-to-end SLA assurance.