Key Insights
- Most operators will make progress toward Level 4 autonomy through individual use cases, but scaling across network domains requires shared data, orchestration, and controls for executing automated decisions. Existing automation remains fragmented across vendor systems, and agentic Artificial Intelligence (AI) cannot resolve these integration challenges on its own.
- 5G-Advanced (5G-A) and 6G will increase the need for coordinated automation as networks support more dynamic services, complex configurations, and AI-driven workloads.
- End-to-end vendors can simplify integration within their own ecosystems, and operators pursuing multi-vendor strategies must coordinate platforms, data, and workflows across suppliers. Open-source integration offers flexibility but leaves operators responsible for maintaining interfaces, security, and reliable coordination as vendors and AI capabilities evolve.
- Agentic AI needs shared context, interoperable interfaces, and clear authorization controls to coordinate actions safely across network domains.
TM Forum describes Level 4 as a highly autonomous network that makes decisions in complex, cross-domain environments using predictive analysis and active closed-loop management. Operators have automated many tasks within individual domains, but their systems often lack the data access and interoperability needed to coordinate decisions across the wider network. ABI Research expects most telecom networks will likely not reach Level 4 in 2027.
How Operators Can Achieve Level 4
Operators typically manage the Radio Access Network (RAN), transport, and core through separate systems. When a transport fault degrades RAN performance and affects customer service, engineers may need to examine several systems to correlate alarms, trace service dependencies, and identify the cause. AI can accelerate this work, but only if it can see data, topology, and service dependencies across domains.
Operators should connect existing automation systems and prioritize use cases such as service assurance and fault resolution. The next step is enabling systems to coordinate corrective actions and verify results. Agentic AI can orchestrate these workflows, but operators need consistent data, interoperable interfaces, and policies defining which actions agents can execute.
What Vendors Are Doing
End-to-end vendors such as Ericsson, Nokia, and Huawei can integrate automation across their own portfolios. This gives operators a more straightforward route to coordinating systems within a single vendor ecosystem, although it does not guarantee Level 4 autonomy across the entire network.
Amdocs, Netcracker, and Mavenir represent another important part of the market. Their software and operational capabilities can help operators automate service assurance, coordinate workflows, and introduce AI into existing operational environments. Mavenir, for example, has outlined a multi-agent service assurance approach intended to detect, diagnose, and resolve issues across multiple domains without replacing existing systems.
The Risks of Building It Yourself
Some operators are building their own automation architectures around open-source frameworks and internally developed integrations. This approach can reduce dependence on a single supplier, provide greater control over agent orchestration, and allow operators to choose the models and tools best suited to their needs.
However, operators also assume responsibility for making the components work together. Open-source agent frameworks do not automatically provide telecom-specific network context, standardized interfaces to every vendor, or safeguards for executing changes in real time. Maintenance, security testing, and auditability also fall to the operator in this scenario.
For most operators, a hybrid model is likely to be more practical, in which the operator owns data models, policy, and accountability while buying orchestration from a supplier.
Is Level 4 Needed for 5G-Advanced and 6G?
Level 4 is not required to launch 5G-A or 6G, but it is increasingly necessary to operate them efficiently. 5G-A introduces more real-time services, including network slicing, differentiated connectivity, and enterprise Service-Level Agreements (SLAs), which add complexities that manual operations cannot operate at scale.
Even more, 6G is likely to add AI-driven workloads and further complexity. Operators that cannot coordinate decisions across domains will therefore find it harder to deliver and monetize premium services.
What Operators Should Do Next
Operators should focus on connecting existing automation systems so they can share network data, and coordinate incidents and corrective actions across domains. They also need clear policies governing which actions can be executed autonomously, with visibility and human oversight. Vendors should demonstrate how their platforms support cross-domain orchestration and interoperability with other suppliers, particularly in multi-vendor environments.
These challenges will be part of the discussion at FutureNet North America on November 4–5. Schedule a meeting with me if you’re attending!
Dimitris Mavrakis