Here are some ways that telcos should be preparing for a future that has everything to play for but is likely to be expensive and demanding to conquer – and some issues to bear in mind
No two operators are the same, nor any two markets. Even so, the common goal for telcos is to identify and capture a sustainable share of the increased value delivered by harnessing AI and network automation.
1. Increasingly agile, adaptive, virtually provisioned and autonomously managed networks will be essential. That does not necessarily mean embedding AI at every level operationally – economics have to come into it. Telcos must look beyond static, rules-based network automation to AI-enabled, dynamic orchestration across multiple network domains – from 5G radio access to IP transport and optical cores.
2. Juniper Research notes that if operators want to achieve zero-touch operation of their mobile networks, this will rely heavily on adoption of agentic AI in the RAN to enable real-time optimisation through reduced latency. The firm advises operators to opt for lightweight models, with a smaller number of parameters, to ensure AI models support lower latency for better customer experience.
3. Zero-touch operations may not make strategic or commercial sense in many cases, especially if the cost is too high, such as in legacy scenarios. TM Forum’s Lupo says AI’s greater value is often in augmenting rather than replacing the work of human experts.
4. As AI’s input increases, so can pressure on peripheral budgets, such as in the supply chain, perhaps in the form of increased Azure storage capacity or soaring spend on platforms like Splunk or unpacking log data. This needs to be factored into the business case and ROI.
5. Preoccupation with achieving ever-better ‘operations’ as an end in itself risks overshadowing delivery of more innovative value for customers. Smarter, more agile network operations will be key to advanced services, but there are other considerations beyond achieving high levels of AN management – see below.
6. On a broader infrastructure scale, AI promises to help optimise resource use in edge data centres in addition to maximising performance. This includes energy use and cooling, which are important to companies’ sustainability efforts and reported carbon footprint. A 2025 report by The World Benchmarking Alliance (WBA) and the International Telecommunication Union (ITU) suggests that the highest-emitting AI systems could potentially produce up to 102.6 million tonnes of CO₂ equivalent (tCO₂e) annually. China Telecom’s evolving strategy, detailed at the 2024 AI for Good Global Summit, is apparently effective at providing mitigation here, targeting better energy efficiency inside and beyond the operator.
7. Partnerships will be crucial to infrastructure-based propositions, as a means to accelerate delivery, share costs and access expertise: AI requires phenomenal investment and there is a global skills gap, and keeping AI skills current is hard given the speed of technologies’ evolution. From strategic tie-ups with AI hardware giants like NVIDIA, to the global cloud giants, innovative app developers and specialist services, telcos must each find the ideal blend of home-grown and outsourced capabilities to achieve their business goals.
8. Real monetisation is about implementing AI to achieve “a positive multiplier effect,” which is eluding most operators, according to TM Forum’s Guy Lupo. He adds, “I see the telecoms industry investing at a level above market rates in getting AI right, yet also failing at a level above market rates,” he says due to following hype rather than, say, heeding how much customers hate some chatbots in assistance use cases. Lupo says sometimes received wisdom in the industry is so strong, nobody likes to question it but attention must be on real monetisation opportunities. The key is for AI to be accountable and dependable so more users and use cases rely on it.
9. Although operators may be feeling overwhelmed by the amount and scope of the work needed to exploit AI and network automation, GSMA Intelligence’s Peter Jarich says, “AI transformation [is] a marathon, not a sprint. As telcos size up their opportunities they should take advantage of emerging frameworks and blueprints created by global industry associations like the GSMA, with input from across the industry.
10. Keep a weather eye on 6G and its implications for network automation and leveraging AI. European telcos risk being left behind because of their reluctance to engage on the topic, whereas Asia is a hotbed of activity. For example, Japanese operator SoftBank’s strategy is built around the integration of AI and communication, and notes that with 6G their integration will advance substantially so AI can optimise communication networks in real-time. Already SoftBank’s AI for RAN is a new architecture integrates AI applications and the RAN on the same computing platform.
This article is slightly adapted from the research report, AI+AN change telecoms’ future, published by Mobile Europe
