From networks to intelligence: How AI is reshaping the telco value chain

Telecoms sits at an inflection point. Data volumes and customer expectations are climbing faster than traditional operating models can keep up, and artificial intelligence (AI) has shifted from a promising R&D topic to an operational and commercial priority. Its impact will be two-fold: driving internal efficiency and opening new markets for digital services and we’re already seeing real-world positive impact on revenue growth, enhanced customer experiences and long-term sustainability, writes George Glass, the chief technology officer of TM Forum.

AI’s first and most visible effect is inside the network. Modern mobile and fixed infrastructures now stretch across 5G, edge sites, massive IoT deployments and, soon, early 6G pilots. Manual planning and rule-based assurance simply cannot keep pace. Hence the rising focus on autonomous networking; software-defined, self-optimising, self-healing infrastructure able to raise alarms, diagnose root causes and, where policies allow, fix faults automatically.

Automate the core

Operators typically climb a ladder of capability. At the augmented rung, engineers use AI tools to search logs or predict capacity constraints, gaining perhaps a 20% productivity lift. The automated rung hands defined tasks – configuration audits, energy management, slice provisioning – to specialist agents. At the top sits the autonomous model in which AI orchestrates traffic and resources continuously, a shift that analysts reckon could absorb 60-80% of today’s manual workload.

This evolution must be led from the top. Boards that treat AI as a side project will struggle to break out of the pilot stage. Progressive executives benchmark maturity, set roadmaps and clarify risk appetite because always-on optimisation touches safety, privacy and brand trust; governance and ethics have to scale with the algorithms.

Democratise intelligence

Operational gains alone will not pay for the networks of the 2030s. Advantage now flows from how widely AI can be deployed across marketing, care, billing and procurement – not just engineering silos. That calls for democratised tooling, domain-specific playbooks and shared data access so that product teams, not only data scientists, can experiment fast.

An AI-native Open Digital Architecture (ODA) is emerging as the reference blueprint. Its principles are straightforward:

  • Data democracy – federated access and common models that let AI learn from customer, service and network domains without copying everything to one data lake.
  • Trust and security – lifecycle management, explainability and policy compliance embedded in every component.
  • Native AI operations – CI/CD pipelines and monitoring for models as well as software.
  • Early adopters such as Vodafone, Telstra and China Mobile report lower integration spend, faster service launches and reduced vendor lock-in.

The external opportunity: AI workloads and the API economy

If internal efficiency explains why AI is unavoidable, external demand explains why it is so valuable. Large generative models thrive on ubiquitous, deterministic connectivity. Telcos own proximity to users and to the edge opens three revenue opportunities.

  • Connectivity as code – initially exposing network features (quality- on-demand, number verification) but the end goal is to move to full network management and operations through standardised vendor agnostic Open APIs.
  • Edge platforms – offering AI inference close to sensors and people where milliseconds matter.
  • Data exchange – sharing anonymised insights that enhance external models while respecting privacy rules.

These models hinge on interoperability. Partnerships with hyperscalers, integrators and software vendors only flourish when onboarding is frictionless, hence the importance of conformance programmes around open APIs.

Reinvent OSS/BSS

Supporting autonomous operations and API-centric business models demands a radical re-think of operations and business support systems. Legacy monoliths trap data, stretch release cycles and inflate what many executives call the integration tax. Composable IT reframes OSS/ BSS and networks as a catalogue of plug-and-play services-product, order, charging, partner settlement and analytics, all invoked through the same Open APIs offered to external developers.

The economic case is compelling:

  • Lower total cost – standard components cut custom interface work.
  • Faster cash conversion – automated order handling shortens lead-to-bill.
  • Strategic freedom – vendors compete on capability, not on lock-in.

Transition requires courage and commitment; integrating legacy platforms can raise costs before savings flow. Smaller operators may favour pre-certified ‘ODA in a Box’ solutions, while larger groups often proceed domain by domain, proving value before tackling core billing.

Partner for a tightly integrated ecosystem

The cumulative effect of AI, open APIs and composable IT is a more tightly woven ecosystem in which value is created by orchestrating capabilities rather than owning every asset. Wholesale fibre retailers, satellite companies, cloud providers and municipalities become both customers and suppliers of digitally brokered services.

Successful orchestration depends on automating partner lifecycle management – from discovery and technical onboarding to revenue – sharing and compliance auditing. Early adopters have shown that removing weeks of manual paperwork unlocks sizeable markets, such as on-demand private networks for enterprise campuses or quality-boost APIs for immersive media.

A call to collective action

AI promises a future where networks are largely self-operating, services are assembled in hours not months and telcos monetise not only bandwidth but also programmable capabilities and trusted data. Realising that promise demands leadership that understands AI’s power and duties, governance that enforces open standards and collaboration that accelerates adoption.

For telcos, the message is clear: embrace AI-native architecture now or risk irrelevance in markets that are rapidly becoming platform-driven and intelligence-centric. For vendors and hyperscalers, openness is no longer optional; it is the price of entry to a value chain that rewards interoperability and punishes lock-in.

AI is more than another technology cycle for telecoms; it is the pivot that will decide which players evolve into orchestrators of the next digital economy and which are relegated to commodity connectivity. The decisions taken today – technical and organisational – will determine on which side of that divide each operator ultimately lands.

George Glass George Glass

Chief Technology Officer