TLDR: Telstra is building its vision for an autonomous network on the foundation of TM Forum’s Open Digital Architecture (ODA), leveraging agentic AI and a new Telstra Reference Architecture Model (TRAM). This strategic shift aims to enhance network self-management, support AI workloads, and deliver personalized customer experiences, with a goal of a highly autonomous network by 2030.
Telstra is aggressively pursuing its vision of an autonomous network, a strategic initiative underpinned by the Open Digital Architecture (ODA) from TM Forum. This ambitious transformation, detailed by Shailin Sehgal, Group Executive, Global Networks and Technology, is set to revolutionize how Telstra manages its infrastructure and delivers services, with a significant emphasis on agentic Artificial Intelligence (AI) and predictive analytics.
The company’s journey towards network autonomy is a continuation of its broader business transformation efforts, including the T22 and T25 strategies. A key outcome of these initiatives is the Telstra Reference Architecture Model (TRAM), which is built on ODA principles and employs an API-first approach to create composable network and technology services.
Telstra’s latest five-year strategy, ‘Connected Future 30’, launched in May, outlines substantial investments in national infrastructure to accommodate the escalating demands of AI workloads. This strategy also focuses on developing user-intent-driven autonomous networks and crafting personalized customer experiences through data and predictive analytics.
Sehgal elaborated on the role of AI agents in autonomous networks, describing them as ‘modular tools that reason and take on specific tasks.’ These agents are designed to be narrow in focus, with individual agents handling functions such as maintenance, optimization, resource management, and support. The vision is for these agents to coexist and collaborate within the network ecosystem.
While acknowledging that the industry is ‘not there yet,’ Sehgal stressed the critical need to define an architecture and deployment pattern for designing and scaling these AI agents. He emphasized the importance of ensuring that the reasoning of agentic AI, as it acts on user intent, is accurate, explainable, predictable, and secure for autonomous operation.
Telstra’s commitment to autonomous networks has already garnered recognition, including an Excellence Award for its innovative approach using an in-house knowledge graph and ‘Knowledge Plane.’ By leveraging ODA, Open APIs, and TM Forum’s best practices for Network-as-a-Service (NaaS) and digital twins, Telstra has moved beyond scripted network service automation to an intent-driven approach.
Telstra International, the global arm of Telstra, is also on a parallel path, aiming for a ‘highly autonomous network’ by 2030. This involves augmenting its subsea cable network to support over 800 Terabits per second (Tbps) of total lit capacity and integrating AI technologies for network performance monitoring. Roary Stasko, CEO of Telstra International, anticipates a threefold increase in traffic by 2030, driven by AI development, necessitating a significant upgrade in network capacity and coverage.
Over the next five years, Telstra International plans to incorporate AI, machine learning (ML), and cognitive computing for network self-management, alongside standard-based APIs to simplify interoperability. A central element of this vision is the creation of a digital twin of its network infrastructure, a virtual replica enabling real-time monitoring, simulation, and optimization. This will allow the network to detect underutilized routes, adjust capacity, and respond to vulnerabilities autonomously.
Mark Sanders, Chief Architect and Head of Autonomous Networks and AI Enablement at Telstra, highlighted that the company’s composable architecture, built on TM Forum’s ODA and a structured telco ontology, is driving its autonomous network strategy. This approach has already yielded tangible benefits, including a 30% reduction in order processing time and a decrease in service creation and onboarding time from 12-18 months to just 3 months.
Also Read:
- Optus Unveils “Expert AI” to Empower Customer Service Teams
- Neo4j Advocates for Data Architecture Shift in Australian Enterprises to Harness Agentic AI Potential
Telstra’s strategic investment in ODA and AI is designed to bridge the gap between network teams and other organizational units, ensuring a common architecture (TRAM) across product, channel, and network experiences. This integrated approach is proving successful in bringing the benefits of Open Digital Architecture to the network side of the business.
