Agile Telco – On Trend July 2026
DTW Ignite 2026 Special
COMMENT
Worried about AI?
– Trust me, I’m a telco
DTW Ignite is always an interesting event, aided in part by its relentless focus on operational transformation at telcos. This means the fragmentation of hyperscale telco events is replaced by targeted topics but the tone has changed in recent years. It has moved from reflecting an industry struggling to preserve profitability to one that is now the core enabler of AI workflows and is itself benefiting from the application of automation and AI. These technologies can transform telcos’ cost bases and, at the same time, position them as flexible providers of essential technologies for a new generation.
If you break it down into a soundbite, attendees are no longer in an industry struggling to survive, they’re engaged in a battle about how to thrive.
That’s no trivial or easy thing. Don’t forget telcos operate regulated, mission critical infrastructure so they can’t roll-out functions, such as network planning or provisioning, as a best effort beta test. The autonomous systems, the AI and the orchestration they adopt must be telco-grade and ready to support modern communication workflows at hyperscale.
I, and others, detected a new awareness of this telco-specific value in Copenhagen. Yes, we’re all still talking about where to find new revenues, how to modernise, innovative methods to increase efficiency and how to ensure investments position telcos with flexibility for whatever comes next but what’s new is the focus on telco strengths.
Trust is now a central telco strength. Industries, enterprises and individuals all recognise how AI and connectivity have already transformed their worlds and know that an even more disruptive wave is coming. That’s not necessarily a negative thing but in that storm, knowing your infrastructure is enabled by a regulated, multi-generational business with a focus on availability, security and quality is an attractive point of differentiation.
Even so, event organiser TM Forum’s own research revealed telcos have a way to go in shoring up that trust. It published a survey uncovering that while 72% of respondents stated that they were confident that their AI was trustworthy, only around one in five of them, just 14% of all telcos, can produce externally reviewable evidence of it.
“Trust me, I’m a telco,” may not have been the words you expected to hear, but they were echoing across the glass roof of the Bella Center at this year’s event.
Enjoy our exclusive coverage in this issue of Agile Telco On Trend.
INTERVIEW
"We’re moving to an intelligent organism driven by a central intelligence from the network to the experience."
Gil Rosen
chief marketing officer, Amdocs
Agile Telco: During DTW Ignite 2026, Amdocs announced advances in Amdocs aOS network workflows including a multi-vendor AI-RAN offering and a blueprint for the next generation of edge AI, analytics and physical AI apps alongside current RAN workloads. How different is this to what has gone before?
Gil Rosen: It’s the foundation for and the reason why we launched aOS. We’re no longer a sector-specific technology provider. For this new era, intelligence needs to be distributed across the entire system and how you operate everything is one of the most important parts of the agentic era. As we move on from being a BSS/OSS vendor it’s about the orchestration of this new ecosystem.
We’re moving to an intelligent organism driven by a central intelligence from the network to the experience across the entire stack. We start from our core competency, which is keeping this operating system open, modular and able to communicate with any vendor in any system. It’s certainly not proprietary.
For us, instead of having the next version of our BSS suite or CES – which we still do – the story is bigger. What used to be operated by people across different technologies is now operated by the technology itself and this is a profound change. For aOS, it’s about trying to reframe from the OSS to the OS story and that’s a significant leap.
Agile Telco: Is the industry ready to take that leap?
GR: If you think of the situation as an analogy for the human body composed of bones, muscles and veins, we used to provide the bones with capabilities in a specific area that did specific things. Today, we’re not replicating the bones, we’re overlaying awareness with a system that has sensing and cognitive capabilities, it’s not what we call deterministic.
Three years ago, we had never heard of a system of record. The industry view was that BSS was ‘the bones’ and it didn’t need agents and AI to help telcos because they’re too expensive. For example, if you go back to the human body, the knee moves up and down. It is a mechanical and repetitive movement that doesn’t need intelligence. It’s not always the knee that needs to move. The recognition came about a year ago that a new overlay is needed so you can drill down into the specific areas you need to operate. The system of record is replaced with an agentic layer on top of it.
That’s necessary because the system of record is not built for scalability and will not be able to serve the agentic layer. You can’t avoid modernising your underlying layer in the long-term, but you don’t have to modernise it right now. An overlay enables companies to manage their transition incrementally, and begin immediately.
Agile Telco: Isn’t there a need for speed here? Telcos can’t move at their usual pace, can they?
GR: Telcos need to realise we are where we are because hyperscalers invested in their technology. It took seven years to develop ChatGPT but we can’t wait another seven years for that level of depth. Now, a lot of strategy around evolving your IT stack has changed and there is a clean slate but no template for telcos to follow.
Every environment is so fundamentally different that you need a custom-made plan for the next five years that delivers quick wins here and there. The underlying network can remain while this happens.
Agile Telco: What are those easy wins?
GR: Having a contract with an enterprise customer is usually very difficult with different service level agreements across different locations. Intelligent automation of that process can reduce some of the most important key performance indicators from order to activation. Today, it might take 12 months and it’s complicated and tedious, so trying to take AI to shorten it to weeks saves resources and the contract has started faster so revenue is generated earlier.
Things like this can be quick wins and we’re seeing a lot of agentic services being launched on the care layer because of that. The network is always behind because the impact of an issue can be a national crisis while a care problem only affects one customer.
Agile Telco: What about autonomous networks, that’s a big theme here?
GR: Autonomous networks are the headline because they’re where the big money is. Network planning is very complicated and encompasses where your customers are and where to lay the next fibre. It also helps people on the business side of the telco who are being empowered to engage with the IT systems using natural language. We have an amazing catalogue product and that has infinite possibilities. It’s like Photoshop, you used to have to be an expert to use it but now anyone can access the benefits.
This creates a completely different level of engagement and employees’ experiences will be profoundly different and that’s very exciting for us.
EVENT REPORT
DTW Ignite – More than a new network, introducing the autonomous telco
Although TM Forum chief executive Nik Willetts stated in his keynote presentation that: “You can’t announce your way to autonomy,” there is a sense that true autonomy did announce its arrival in the telco industry at DTW Ignite 2026. ‘The autonomous telco’ is an umbrella term that covers the new wave of autonomy poised to transform telco operations over the coming decade.
It’s important not to confuse autonomous telco with autonomous networks, they’re different things. The autonomous network forms just one part of the autonomous telco and will enable network operations to become truly autonomous, but the full picture is of a telco that operates autonomously across the network, OSS/BSS, customer experience and its own business processes.
It’s also important not to conflate AI and autonomous networks. The two are interlinked and often deploying hand-in-hand so there’s an understandable tendency to bring the two together which reflects likely deployment scenarios and the development of the AI/AN ecosystem.
That full picture of a completely autonomous telco is a bigger goal than discrete automation projects, targeted application of AI and isolated autonomous pilot projects. Those are all happening and they all contribute to the eventual achievement of an autonomous telco but they’re stations on the journey, not the final destination.
All change
“Customers are still thinking about: how do I reinvent myself?” confirms Patrik Eriksson, a vice president and the head of Telco Industry EMEA at Oracle. “They’re moving from technology providers to digital companies but it’s very basic in many ways. Now, they’re turning to outcome-driven discussions about growth, efficiency, modernisation and trust.”
“AI changes everything – it’s a bold statement but we genuinely mean that,” he adds. “The change can’t be in the network only or just the BSS or the operations layer, it has to follow the entire business. The autonomous network is a sub-component of the autonomous telco.”
Anita Döhler, the chief executive of NGMN Alliance, a global, operator-driven industry alliance, sounds a more cautious note. “I don’t think agentic AI is an either/or for 6G but our operator members are very keen to get the basis for autonomous networks right,” she says. “We have had working groups in previous years focusing on what it means to be cloud-native because that is seen as a basis to move to fully autonomous networks.”
It’s important to understand the scale of the tasks facing operators. “Moving to an agentic AI-enabled operating model is complex,” Döhler adds. “Even if you looked at what is needed for operators to operate cloud-native networks, there’s a need to upskill people, embrace new technologies and processes. NGMN has published specific guidance targeting MNOs around how to move to agentic AI based operating models, suggesting five levels.”
Extract early business benefits
The business is where the change originates from as telcos look to innovate and offer compelling experiences to their customers. Alex Bloomfield, the R&D director at Lifecycle Software, sees the application of AI helping telcos focus less on a single prize and more on complimentary services. He cites an example of a mobile operator in the Channel Islands that is achieving differentiation through adding value. It is offering 10% off vouchers at a local supermarket and has seen a doubling of customers porting to it in the first two months of the campaign.
Bloomfield says the company’s new agentic AI framework, Nexus IQ, allows users to build their own workflows across thousands of data points. “This allows us to be even more personalised and find behaviour we didn’t even know about before,” he explains. “This increases the opportunity to interact with customers with micro-moments. It’s about being targeted, we’ve built a framework that enables us to ingest data and much better understand the behaviour.”
“We have
one example use case that has seen a 33% increase in add-ons being purchased,”
he says. “That’s a genuine return on investment and it makes moving into the
agentic AI world unavoidable. The barrier to entry with AI is lower than it has
ever been.”
Foster familiarity
Targeting specific areas of the telco business is not only achieving valuable results, it’s fostering familiarity with AI, how it can be used and what is realistic for it to be applied to. Tom Cox, the founder and chief executive of Humara, specialises in creating virtual sales agents that outperform humans. The company, which focuses exclusively on the telecoms market, has used learnings from human telco agents to educate its virtual agent and is now looking to expand into Europe, taking it beyond its roots in the UK and US markets.
“The problem we solve isn’t really a telco-specific problem, it’s when someone is making a meaningful purchase and they have a confidence threshold to overcome,” he explains. “A telco contract can be really hard to keep people above the signing threshold. Our job is to identify and solve every point of friction and mirror the techniques of the best salespeople.”
Humara is so confident of its ability to keep customers engaged throughout the sales process that it charges based largely on performance. “Humara sells by building trust and rapport, the next step is discovery,” he adds. “Historically, the product was quite hard to onboard and it could take months to do so, but we can now onboard it in weeks so smaller telcos and MVNOs can be added.”
Beyond the point
Point solutions are one thing but taking a discrete functional area of operations takes AI adoption beyond the single issue. “The question we see now is what can I do with AI?” confirms Jeffrey Spiess, the product director at Motive. “Virtual assistants are replacing scripted chatbots. The value we add is to expose the capabilities of our system to an agentic AI virtual assistant independently of the large language model of choice.”
Drawing on its device management heritage, Spiess also sees Motive taking agentic AI and applying it to autonomous and predictive remediation of devices. “We always could do orchestration of remediation of an issue,” he explains. “For example, if the right fix is to reboot a home gateway, we and do it at a friendly time but now we’ve added the ability to look at the data and predict when a customer is likely to experience the issue.”
“We’re definitely using agentic AI but a lot of use cases don’t require AI,” he adds. “There is the cost and latency of AI to consider and there are still some things to be done with machine learning. With AI you still have to train a model, you can do it with LLM but you’re going to have to pay for fibre connectivity. It’s better to tunnel it and narrow the application down into an issue.”
DTW Ignite 2026 by the numbers
Source: TM Forum
Attendees at DTW Ignite 2026
0
of telcos plan to achieve Level 4+ autonomy within three to five years
0%
of telcos believe their AI is trustworthy but only 14% can produce evidence to prove it
0%
Results and returns
That focused targeting of when to apply AI is sensible today as telcos continue to experiment. They also need to ensure early deployments generate results and returns. These are necessary steps to build familiarity and confidence and at no point should telcos seek to boil the ocean with an everything-AI approach. However, they need to do more than confine their attention to point solutions. Ideally, AI needs to be able to be applied across the entire telco stack, enabling value to be contributed wherever and whenever practical.
“It’s not new, we always had the ecosystem to consider and, as with any relatively new opportunity, it’s taking time,” Eriksson says. “We’re trying to contribute and service providers have a vital role to play in the ecosystem because they have the final business outcome overall. We need to move from AI being a technology journey, it has to become an outcome journey owned by the business people.”
Eriksson cited a presentation at the show by the CIO of KPN, Paul Bosch, which described how the telco is inspiring penetration of AI to fix the network and create an intelligent layer. “There is a move to the business people within telcos to help with that discussion,” Eriksson explains. “What I’m seeing is the successful ones are taking point solutions, not horizontally but vertically.”
“Seeing is believing, simple use cases include revenue leakage and quality assurance and good things are happing with those from an AI perspective with customers,” adds Eriksson. “These are simple use cases with an immediate profit and loss impact.”
Verticals unlimited
Eriksson advises avoiding the trap of looking at all AI in the same way. “There are two types of AI – horizontal and vertical,” he says. “There is horizontal AI in which a simple query might result in reducing your costs but it’s not very sophisticated. In contrast, the vertical is how to look at the entire process. For example, looking at order to cash and implementing AI into the overall workflow.”
“In that scenario, AI can look, predict, recommend and take action automatically,” he adds. “It can help time to market, profit and loss, and speed of operations if it is in the workflow. Horizontal AI is the foundation but vertical AI is starting to happen.”
This remains a giant leap for telcos. “We have the full stack – the infrastructure, the data and the applications – so it’s very easy for us to imagine how AI is changing everything from the network to the management layer,” Eriksson explains. “There are a lot of discussions around why this isn’t happening and those break down into a few primary reasons.”
Eriksson cites insufficient focus on having good data to feed successful AI. It’s the old database paradigm of ‘rubbish in, rubbish out’ but telcos need to ensure they have a consistent data position to ensure AI can run effectively. In addition, telcos need to clean up and streamline their legacy. They can’t expect to succeed in the autonomous era if they force traditional operational complexities and inefficiencies into the new model.
EDITORS TAKE
Is there method behind the AI madness?
Have your say
Give us your opinion on AI in telecoms, automation and whether the industry is finally replacing hype with method. Ring in and leave a message or use email to comment.