After years of stagnant development, can AI provide telcos with the tools to forge new opportunities and drive sustainable revenue growth?
While a wider ecosystem around AI is taking root, the telco industry is arguably one of the hardest environments for AI to transform. A striking gap, still remains in AI adoption across Asia-Pacific’s industry, with a TM Forum study of telecom tech leaders finding that while nearly nine in 10 operators recognise AI’s importance, very few are putting it to work in their core operations.
The study warns that APAC operators’ intentions for AI to reshape how technology creates customer value are at risk of stalling at execution. The challenge now is to move from automation to autonomy at scale before disconnected pilots add cost and complexity rather than delivering growth.
TM Forum is a global alliance of more than 800 organisations across the connectivity ecosystem. Members include the global top 10 communication service providers, leading hyperscalers, network equipment providers, and a wide range of vendors, consultancies, and system integrators.
To get a feel of the ground in APAC and understand the challenges faced by telcos in their AI adoption, iTNews Asia speaks with George Glass, CTO, TM Forum.
In his role, Glass enables digital service providers to simplify the transformation of their IT and network estate, with the aim of helping them drive AI and digital revolution that is happening across all industries today.
iTNews Asia: Telcos have traditionally been slow to innovate due to a business environment that is more risk-averse, low R&D spending and legacy infrastructure/old operating models.
They have also struggled to monetise their high capacity 5G networks, while device, app and cloud providers have benefitted.
Moving into 2027, will this landscape really change? How do you see telcos in APAC evolving?
As we move into 2027, we’ll start to see separation between the operators that are genuinely changing how they run their businesses and those simply adding new technology onto old operating models.
In APAC, we’re already working with telcos that are moving from automating individual processes to autonomy at scale across multiple network domains simultaneously. Momentum is building and research is showing that Asian operators report some of the highest average autonomy scores worldwide.
Many operators have spent years building faster networks without necessarily creating faster growth. 5G demonstrated that adding capacity alone does not create a new business model.
AI presents another opportunity, but operators cannot afford to make the same mistake by simply bolting it onto existing systems. Those that pull ahead will use AI and automation to change how they operate and create services around outcomes customers will actually pay for.
iTNews Asia: How are they currently using AI? What new capabilities can AI provide them?
AI can give operators the ability to move beyond simply identifying network problems to analysing, deciding and acting on them autonomously. Fault management is a good example, because identifying a problem is only the beginning as someone still has to work out the root cause, decide what action is required and safe, and then fix it.
Japan’s Rakuten Mobile has demonstrated that this whole loop can become autonomous on a live commercial network. The mobile network operator’s system can detect and analyse faults, determine the appropriate response, execute the remediation, and check that service has recovered, without routine human intervention.
That changes the role of the engineer. Instead of repeatedly firefighting known problems, engineers can focus on improving how the network performs and what it can deliver.
We are also seeing autonomy applied to energy efficiency, service assurance and network optimisation, but we should be much less interested in collecting hundreds of AI use cases than in whether operators can make successful ones repeatable.
The real prize comes when autonomy stops being a collection of clever projects and becomes part of how the network operates.
iTNews Asia: Your recent study with telecom tech leaders shows that while the majority see AI as critical, only a very small two percent have successfully integrated AI systems that can carry out real tasks within their day-to-day operations.
Why do you think this is so?
Telcos do not have an AI experimentation problem. They have a scaling problem. What makes this particularly interesting in Asia is that we can see both realities at once. There are highly advanced deployments demonstrating what autonomy can achieve, while much of the wider industry is still working out how to industrialise AI.
Closing that gap is what the next phase is about.
iTNews Asia: What challenges do they face?
The difficulty is that operators are introducing AI into decades of technology, processes and operational knowledge built around human decision-making.
Almost half of respondents in our study cite skills as a major obstacle, followed by poor data quality, cost and concerns about AI accuracy or hallucinations.
More than a third say critical human decision-making is not sufficiently documented, making it much harder to hand decisions safely to AI.
iTNews Asia: What can they do to overcome the hurdles they face?
Overcoming that requires more than another round of pilots. Operators need better data, modern and modular systems, clear governance over what AI can decide, and processes designed around autonomy rather than simply automating the way things have always been done.
Ultimately, we need to have trustworthy AI, a concept being developed by TM Forum and its members to address these very issues.
iTNews Asia: How critical will AI be in turning their ambitions to reality? Do they need to change their business models to stay viable and competitive?
AI is absolutely a critical enabler, but it isn't the business model.
If operators use AI simply to do the same things faster and cheaper, they may improve some margins, but they won't fundamentally change their position in the value chain. The bigger opportunity is to make networks responsive enough to sell outcomes that customers will pay for, rather than just connectivity.
An autonomous network that can respond to changing demand, adjust resources, and manage service performance in real time, makes services possible that would otherwise be too costly or complex to deliver manually.
We are already seeing signs of the commercial potential. Telstra, for example, has reported 30 percent faster time to market and a 20 percent reduction in operating expenditure through 5G network slicing, helping open the door to new services and offerings.
I see autonomy as a growth agenda, even more than as an efficiency agenda. AI will be critical, but it isn't a magic bullet. Operators still need a plan to change how they operate and what they sell.
The competitive advantage will come from turning network intelligence into services and outcomes that customers actually value.
Can AI be the panacea to jumpstart telco’s growth? Which areas can they make money from?
If we treat AI as a panacea, we'll repeat the mistake the industry has made with previous technology cycles and assume the technology itself will create growth. AI can change the economics of running a network and make services viable that previously required too much manual intervention, but somebody still has to identify a customer opportunity worth creating.
- George Glass, CTO, TM Forum
We can already see the operational case taking shape. Research found that areas such as energy-efficiency optimisation and service assurance are averaging autonomy scores of 3.9 out of 5 among the operators we assessed. The opportunity now is to turn that operational progress into commercial value.
That could mean providing an enterprise with a connectivity outcome that is configured, assured and adjusted dynamically instead of selling a fixed product, or exposing network capabilities so partners can build them directly into their own services.
AI creates the conditions to do those things at a speed and scale that would be impossible to achieve manually.
So, the question goes beyond where telcos can make money from AI. It's what new services become commercially viable when AI and autonomy remove the operational complexity that previously stood in the way.
