Practical AI-Powered Network Innovations at DTW Ignite 2025

TM Forum’s annual Digital Transformation World (DTW) Ignite event is more than a talking shop. Unlike most other industry events, a major part of DTW Ignite are the practical efforts to build future telecom networks. Alongside the conference presentation and panels, there are numerous collaborations involving multiple companies, as the industry together builds working prototypes that could be extended into commercial offerings.

This year, three main streams dominated the show. Firstly, the shift to open digital architectures (ODA) and composable IT and ecosystems. The industry’s goal is to simplify operations with zero-touch approaches that speed innovation and raise the quality of customer experience.

As everywhere else in 2025, AI dominated the other two streams: AI and data, where analytics is being transformed using classical machine learning, with generative AI helping to organize unstructured data. Unlike many industries, the scale of data creation in telecom networks is already vast, and high levels of reliability are also essential.

The shift to autonomous networks is perhaps the most high-profile initiative of TM Forum. The goal is simple: automate to improve efficiency and help operators manage increasingly complicated networks, without needing a vast increase in the number of expert network operations technicians.

The parallel TM Forum continues to strike is with the automotive industry’s levels of autonomous driving. Over the past few years, most operator networks have fallen somewhere between level 2 and level 4, depending on the part of the network and the operator concerned. But many of these grades have proved controversial. Cultural differences between operators have meant some have overembellished their achievements, while in other markets, operators have taken a conservative view and understated their progress.

The benefits operators aim at with their autonomous networks are significant, including: speeding innovation, allowing operators to deploy more new network capabilities; reducing manual effort, so lowering risk of human error; increasing reliability, as networks automatically detect and resolve problems, or predict and avoid faults before they happen; lowering operational costs, by fault avoidance and requiring fewer salaried experts.

But without a standardized benchmark, it’s hard for leadership teams to manage network teams, correctly forecast cost savings, or know if their network is achieving the benefits it should at a given autonomous level.

Standardizing Autonomous Network Level Assessments

This year, TM Forum has pioneered the Autonomous Network Level Assessment Validation (ANLAV) service to provide an independent benchmark. This builds on the existing Autonomous Network Level Evaluation Tool (ANLET). At DTW Ignite, TM Forum picked out four leaders among the more than 30 communication service providers that have assessed their networks: Denmark’s TDC secured validation for its radio access network (RAN) energy efficiency optimization; Telefonica’s Vivo claimed efficiency gains of 90% in route convergence and 60% in budget approval time, having reached autonomous network level 4 in network creation and planning for its transmission network.

Two Chinese operators were also highlighted, emphasizing again China’s role on the leading edge of telecom network progress. China Mobile reported an 80% reduction in major network faults with an associated electricity saving of 7 billion kilowatt hours as it approaches level 4 in several domains, such as IP backhaul and RAN fault management. It scored 3.8 in wireless fault scenarios and 3.6 in core network faults. Similarly, China Telecom is closing on level 4 this year, and claims it has already saved 1 billion kilowatt hours.

Deepening Network API Collaboration

There was another step on the path to enterprise-accessible APIs, as the GSMA and TM Forum made a joint announcement of collaboration to create a single certification programme for APIs. Operators will be able to ensure their API implementations conform with TM Forum Operate APIs, Camara Service APIs and Open Gateway. This highlights the complexities of exposing APIs to third parties and why many network equipment vendors have their own initiatives to aid operators with API implementation and commercialization.

Building the Future with Catalysts at DTW Ignite

One of the major features of the DTW event was the Catalyst showcase, spread across three areas. Catalysts are collaborations involving a group of companies, usually including several vendors and at least one mobile operator. Their goal is to build a working prototype or proof of concept that demonstrates innovation and network transformation. There were three that caught our eye this year.

1. “AI-enabled system for vehicle-road-cloud collaboration — phase II”

This effort won the Best Moonshot Catalyst: Attendees’ Choice Award. It covered core network capabilities as well as wireless innovation, AI and edge computing, with the goal of improving safety and traffic efficiency in urban areas. Notably, this is a cross-domain, end-to-end project that’s increasingly needed not only for automotive but also as part of autonomous networks and improved operations and maintenance innovation. The three participants were China Telecom Fufu Information Technology, Primforce and ZTE.

2. “Satcom with an edge — phase III”

Here the goal was to offer seamless and on-demand access to multiple non-terrestrial networks to complement existing cellular connectivity. The main use demonstrated was to create a software-defined wide area network for emergency responders when the cellular network is unavailable. The system is able to intelligently choose the right network for the performance needed, handle service assurance end-to-end, and bill different suppliers for each part of the connectivity solution. Participants were: Alvatross, Amartus, BolgiaTen, Celfocus, Digital Route, Enghouse Networks, Kratos, Oracle and Radcom.

3. “L4 autonomous networks: agent-powered zero-touch workflows”

Like many other efforts this year, this Catalyst was cross-domain and aimed to improve the end-to-end experience. There were two parts: a fault handling agent that provides real-time network awareness and automated response to speed fault resolution; a wireless network optimization agent that makes automatic adjustments based on live network conditions to improve signal quality. This project highlights the role of agentic AI to advance autonomous network evolution. Participants included: AsiaInfo, Ultrapower, Huawei, Inspur, ZZ and ZTE.

Cross-Domain Network Management Is Becoming Essential

Operations teams increasingly need to be able to assure services end-to-end, especially with the rise in network slicing, and to resolve faults that could appear anywhere in the network. Networks also threaten to become even more complicated with the integration of non-terrestrial networks in recent 5G standards, and looking ahead to the “ubiquitous connectivity” goal of 6G, operators need to move beyond network silos.

An example of such cross-domain solutions is the AIR Net framework for end-to-end network automation from ZTE. It includes multiple large AI model agents and tools to increase network awareness, improve analysis and enable adaptive service orchestration.

At DTW Ignite, the industry sees AI — both traditional machine learning and generative AI — as vital to move beyond siloed views and create a flexible, automated, responsive, reliable and innovative network. The term “AI native” is increasingly being used to signify this movement. “Carriers used to manage networks”, explained T-Mobile US’s president of technology, Ulf Ewaldsson, “Now we must manage intent, thanks to the ingredients of AI, to provide the best possible customer experience at any given moment”. And, by using AI tools and shifting to automation across network domains, telecom operators can reshape their networks for current and future needs.

Written by:
Posted on July 4, 2025
Share