- 5G Standalone gives operators new ways to differentiate services through slicing, private 5G and connectivity with defined service levels
- Ericsson says carriers should pursue autonomy domain by domain, rather than expecting to make an entire network fully autonomous at once
- Looking toward 6G, Ericsson’s Lucky La Riccia expects AI to be native to the network, increasing pressure on operators to rethink how they monetize services, SLAs and differentiated experiences
ERICSSON OSS/BSS SUMMIT 2026, ZURICH, SWITZERLAND – Operators chasing autonomous networks shouldn’t let messy legacy data become an excuse for standing still.
That’s the view of Ericsson’s Lucky La Riccia, head of cloud software and services for EMEA, who argued that carriers can make progress toward autonomous operations even when their underlying data is less than perfect. The trick is putting guardrails around that data and concentrating first on network domains where greater autonomy produces a clear business benefit.
“Maybe it’s okay to have some unstructured data. Maybe it’s okay to have data that’s not perfect, or you can say a little bit imperfect, as long as you can build guardrails and tolerance around that,” La Riccia told Fierce Network in an interview at Ericsson’s OSS/BSS Summit in Zurich.
That matters because operators are dealing with networks containing thousands of physical assets as well as legacy IT systems, making perfectly clean and current data difficult to achieve. Rather than wait for perfection, La Riccia said operators need to understand the limitations of their data and establish appropriate controls around how autonomous systems use it.
Ericsson is also adjusting the familiar “human in the loop” concept. La Riccia described a “manager in the loop” model in which a person can manage an agent, stop an action, review it and allow it to proceed when appropriate.
Autonomy won’t happen everywhere at once
Just as operators have different data challenges, they’re unlikely to follow the same path toward autonomy.
La Riccia pointed to different approaches already being taken by operators in EMEA. Telia’s Nordic operations, for example, are starting with a common network inventory view so the operator has a consistent view of network assets for orchestration and provisioning. Swisscom, meanwhile, is starting with a cloud-native core and an intelligent automation platform, including closed-loop automation in the radio access network.
The underlying idea is to identify a high-value scenario first.
“Each customer kind of identifies what’s my high value scenario, where do I want to introduce autonomy, and what’s the value it will give me commensurate to the investment I want to make,” La Riccia said.
Ericsson therefore isn’t expecting operators to suddenly flip a switch and make their networks fully autonomous. Over the next year, La Riccia said the company wants to see more customers move individual network domains higher up the autonomy ladder.
“Each customer will not get to level four autonomy across a network in the next year,” he said. Instead, operators may select particular domains and decide “how do I get each of those domains up to a really good level that helps me run my business?”
The return doesn’t necessarily have to come from lower operating costs either. La Riccia said greater autonomy could create value by helping operators bring services to market faster, improve productivity or obtain customer insights they couldn’t previously access.
5G Standalone creates the service opportunity
Autonomy becomes more important as operators move deeper into 5G Standalone (SA) and attempt to create differentiated services instead of relying on traditional best-effort connectivity.
Slicing is part of that opportunity, but La Riccia cautioned against viewing it as the entire story. “It doesn’t necessarily have to be a slice; [it] can be private wireless and private 5G, as an example,” he said.
The bigger opportunity is to provide connectivity tied to a particular service level, including for emerging physical AI applications such as robotics. Those workloads may need mobility and connectivity at a specific location or at the edge, capabilities that a data center alone cannot always provide, La Riccia said.
Once an operator provisions a differentiated service, autonomous operations can help monitor whether it is performing correctly, identify developing network problems and potentially take corrective action without human intervention while protecting the customer’s SLA.
If operators are going to sell differentiated connectivity, they need a network capable of continually delivering what they have promised, he said.
From adding AI to 5G to AI-native 6G
That challenge becomes even more significant as the industry starts looking beyond 5G. La Riccia drew a distinction between today's efforts to introduce AI and autonomy into 5G networks and the expected architecture of 6G.
“What we’re doing in 5G is building autonomy and introducing AI into networks. 6G is AI native at the core,” he said.
That transition could also force a rethink of telecom monetization. Rather than charging around familiar connectivity plans, operators may increasingly need the ability to monetize specific services, SLAs or differentiated customer experiences.
“I need to monetize a service, not necessarily, or an SLA, or I need to monetize a different, unique experience that’s differentiated in a particular way,” La Riccia said. “So, you need to get all your monetization layers ready for that.”
For operators, that makes the road to 6G about more than another radio upgrade. The work taking place now around autonomous operations, 5G SA and new monetization models is laying groundwork for networks in which AI is embedded much more deeply into how services are operated and sold, La Riccia concluded.
Read more about Ericsson and 6G on Fierce Network:
Verizon, Ericsson dial up 6G in 7 GHz field test
T-Mobile cites AI-RAN gains with Ericsson in NYC