Ericsson has partnered with Australian telecom company Telstra on an autonomous network initiative.

The collaboration will center around a technology unit for experimentation and design, with a view to assessing the reliability of AI data models. The pair also aim to demonstrate and validate solutions to challenges arising from multi-party and multi-domain environments.

Using Ericsson offerings, both firms plan to advance intent translation frameworks to connect business goals with network actions. Key to this is the build-out of a knowledge plane consisting of a data and AI information layer, which can monitor, analyze, and manage the network on behalf of network teams.

On the AI front, Ericsson and Telstra said they would foster explainable AI, with the intent of making AI-driven decision-making transparent and comprehensible to users.

The deal builds on previous collaborations between the pair, such as the deployment of what was claimed to be the first programmable, 5G Advanced-native network in Asia-Pacific (APAC). It also expands Ericsson’s autonomous ventures in APAC, following a similar arrangement with Malaysia’s CelComDigi from April, as well as its extensive work to date with Malaysian telecom operator, Digital Nasional Berhad (DNB).

Ericsson and explainable AI

In conversation with SDxCentral at Ericsson’s OSS/BSS Summit 2025 event, Ericsson's head of BOS portfolio, Jason Keane, said the partnership was unique amongst its recent autonomous network team-ups due to its exploratory nature.

“Together, we're not just looking at a single kind of use [as] we want to build this knowledge plane to represent our whole network. We have multiple data sources, multiple orchestrations. What does this look like when you grow it up?”

Keane also delved into the importance of trustworthy AI for customers, saying it went beyond simple guarantees against machine-generated hallucinations.

“Let's just say something goes wrong. You can't just say, ‘We've no clue, it was the AI.’ That unaccountability will never wash. You have to explain what went wrong and why the system made a mistake.”

“With trustworthy AI, we can trust its outcome and secure its outcome in line with the best interests of the business,” said Keane.

Ericsson's autonomous footprint

The Telstra partnership is the latest in Ericsson’s ventures into autonomous networks, which aim to replace current rule-based automation with AI-driven operations capable of automatic optimization, configuration, self-healing, and enabling network actions with minimal human intervention.

The firm’s AI-centric Cognitive Network Solutions (CNS) unit announced a collaboration this year with Amazon Web Services (AWS) to drive autonomous network development using agentic AI.

The European giant was also selected by Orange Group this summer for 5G network slicing orchestration across Europe, touted by Orange as a crucial step towards autonomous networks.

In June, Ericsson claimed to have achieved the first global verification for Level 4 (L4) autonomy of operations in a specific scenario according to TM Forum standards, as pertaining to its Predictive Cell Energy Management (PCEM) solution.

Away from Ericsson, recent autonomous initiatives have come from IBM and its recent IBM Network Intelligence release. An artificial intelligence for IT operations (AIOps) platform from Ribbon Communications, meanwhile, was designed to help service providers and enterprises transition to autonomous networks.