Telstra & Ericsson expand push on autonomous networks

Telstra and Ericsson have expanded their collaboration on autonomous networks and will co-lead a new industry working group on network knowledge and trust.

The initiative addresses a key challenge as the telecoms industry moves towards more autonomous operations: how systems across different vendors and network domains can share information in a consistent structure and with enough trust to support automated decisions.

Under the expanded arrangement, the companies will focus on three areas: developing a common taxonomy for network knowledge, examining ways to exchange and combine that knowledge across systems and domains, and studying how autonomous AI agents can be grounded in trusted knowledge and safety guardrails.

The work builds on an earlier collaboration framework that examined business-led use cases, proposed solutions and practical lessons for the telecoms sector. This next phase shifts the focus to the governance, trust and knowledge structures needed for higher levels of network autonomy.

Knowledge problem

At the centre of the effort is the idea of a Knowledge Plane, described by the companies as a formal representation of what a network is, what it should be and what it is expected to do. This includes engineering rules, policies and safety guardrails that both people and automated systems can use when making operational decisions.

As telecoms operators and equipment suppliers develop their own approaches to that layer, differences in terminology, structure and trust levels can create silos. These can make it harder to combine information from multiple sources, preserve its provenance and assess its reliability before action is taken.

That matters because autonomous networks rely on software and AI systems to monitor conditions, analyse events and carry out changes with less direct human intervention. If the underlying knowledge is inconsistent or cannot be exchanged easily, operators may struggle to deploy automation across complex environments spanning multiple domains and suppliers.

The joint working group will examine how users and systems can determine the level of confidence to place in particular knowledge before acting on it. It will also look at patterns for combining knowledge across vendor sources, structural formats and time while preserving lineage and trust.

Another strand of the work concerns what the companies call knowledge-grounded agency. In practice, this means exploring how autonomous agents can use trusted knowledge throughout the full cycle of intent, awareness, analysis, decision and execution.

Industry angle

The working group's findings will be shared with industry bodies including TM Forum. The aim is to feed the results into broader standards discussions and encourage wider adoption of common approaches across the sector.

This reflects a wider issue in telecoms, where automation efforts often progress within individual networks or vendor ecosystems but are harder to scale across mixed environments. Interoperability has long been a sticking point in network management, and the shift towards AI-led operations adds another layer of complexity around governance and trust.

Mark Sanders, Chief Architect, Telstra, said: "Autonomous networks require more than the automation of individual tasks. They depend on a shared understanding of what the network is, what it is expected to do and how much confidence can be placed in the knowledge supporting each decision. By working with Ericsson on common approaches to knowledge and trusted agency, we can help address these foundational challenges and contribute to the industry's progress towards more interoperable and dependable network autonomy."

Ericsson positioned the project as part of a broader industry effort rather than a closed bilateral initiative. It said common methods for managing knowledge and trustworthy AI will be needed if autonomous functions are to operate consistently across different technical environments.

Mats Karlsson, Head of Solution Area Business & Operations Support Systems, Ericsson, said: "The evolution towards autonomous networks will require collaboration across the telecommunications industry. Common approaches to knowledge, knowledge exchange and trustworthy AI will be important to enable autonomous capabilities to operate more consistently across domains and technology environments. Our expanded collaboration with Telstra is an opportunity to explore these foundations, share outcomes with the industry and support the development of more open and scalable approaches to network autonomy."