Telstra's tech chief asks telcos to fix basics before chasing AI

You have probably sat in a strategy session where someone sketched a future of autonomous networks, intent-based slicing and AI-powered operations while the room nodded along. Then you walked back to your desk and opened a ticket in a system that still requires three manual handoffs. Kim Krogh Andersen, Group Executive for Network, Product and Technology at Telstra, stood in front of telecom peers at the FutureNet Asia event and named that gap. Telecom service providers are not ready for the AI economy, he said, and the problem is not vision. The problem is foundations.

Andersen shared Telstra's own position. A recent internal assessment placed the company below Level 2 on TM Forum's Autonomous Networks Framework, a six-level scale that ranks progress toward full network autonomy. "It is a bit of a downer to talk about intent-based slicing before you get the fundamentals right," he said. He argued the industry will not reach Level 4 without letting go of legacy systems and simplifying networks and operations first. "AI exploits the vulnerabilities, the basic stuff we have not done in the past," he added, pointing to frontier models that can find and exploit gaps at machine speed.

What this means for infrastructure leaders

Andersen's candour is useful because it reframes the AI conversation. Most telcos are selling connectivity in units, buckets of megabytes and gigabytes, not outcomes, he said. The industry should adopt a product-engineering mindset instead. "No one on Earth can convince me that compute and storage in a rack in the data center is more sophisticated than connectivity," he argued. His strategy is to get back to the core, then grow and innovate from it.

That shift requires treating the network as a product. "The network is our product, and we need to treat it that way," Andersen said. "That means adopting a product-engineering mindset, continuously improving network capabilities and managing them throughout their lifecycle, rather than deploying equipment and revisiting it only when it reaches end of support." It also means addressing legacy infrastructure, because you cannot build the level of automation and intelligence required for an AI-driven future on fragmented and outdated foundations.

The timeline problem

Andersen was explicit about urgency. "We cannot wait five or 10 years to prepare for it, we need to build the foundations now," he said. He described the network as always-on connectivity supporting agents, devices and autonomous workflows across society. In an AI-driven world, connectivity will underpin far more than communication. That vision only works if the infrastructure can handle machine-speed decision-making and if the operations stack is not held together by manual workarounds.

The TM Forum framework Andersen cited is a maturity model. Level 0 is manual operations. Level 5 is full autonomy. Telstra's self-assessment put the company below Level 2, which means most processes still require human intervention. Andersen's point is that the industry is talking about Level 4 capabilities (intent-based orchestration, closed-loop automation) while operating at Level 1 or 2. The gap is not a technology problem. It is a legacy problem, a process problem and a prioritisation problem.

What L&D and innovation leaders should take from this

Andersen's comments are a reminder that AI readiness is not about model access or pilot projects. It is about whether your infrastructure can support the workflows AI enables. If your network or operations stack still depends on manual handoffs, you are building on sand. The same logic applies to learning and development. You can train teams on prompt engineering and agent design, but if the systems they work with cannot close the loop, the training will not scale.

The product-engineering mindset Andersen described is also relevant beyond telco. Treating your network, your platform or your capability stack as a product means continuous improvement, lifecycle management and a focus on outcomes rather than features. That shift requires leadership that is willing to name the gap, as Andersen did, and invest in the basics before chasing the next frontier.

Sources:

• Light Reading, October 6, 2026