The management question is not whether CSPs can add AI, but how they decide where AI belongs first. For telecom leaders, the sharper issue is portfolio discipline: which AI initiatives are truly strategic, which are just efficiency theatre, and which should wait until data, process ownership and controls are mature. Without that filter, AI programmes can fragment into disconnected pilots across network, care, marketing and IT.
That makes governance the real differentiator. AI in a telco touches customer experience, regulated data, resilience, and network integrity at the same time. Leaders need clear decision rights for model approval, risk acceptance, vendor selection and escalation when AI recommendations conflict with operational policy. The practical test is whether the organisation can run AI as a managed service with accountable owners, not as a collection of experiments sponsored by enthusiastic teams.
The second trade-off is speed versus control. Moving AI into production may improve automation and service quality, but it also increases exposure to model drift, hidden technical debt and dependency on external platforms or GPU supply. CIOs and transformation leaders should ask: Which outcomes justify the operational complexity? Which processes can tolerate partial automation? How will we measure value beyond cost take-out?
Workforce change is the final constraint. Upskilling is not just a training issue; it is a redesign of roles, handoffs and accountability between network engineers, data teams, operations and procurement. The management implication is that AI programmes should be funded alongside operating-model changes, not after them, or the organisation will automate faster than it can adapt.
Global AI adoption presents communication service providers (CSPs) with a rare opportunity to transform and unlock new growth. In my latest report, The Dawn Of AI-Powered Telcos, I examine how leading telecom operators are responding to current trends by entering the AI infrastructure race and becoming AIโpowered from the core. Theyโre achieving this by pushing AI into their network, engineering, IT, business, service, and marketing operations. The report unpacks these developments and provides rich examples from operators in various regions.
Picks,ย Shovels, Andย Wires: CSPs Making A Big Play In AI Infrastructure On Their Terms
Many CSPs admit having missed prior opportunities during the rise of OTT (e.g., streaming services) and cloud computing and feel concerned that their traditional value proposition anchored in network connectivity (e.g., voice and data) is too commoditized. Thatโs not something they want to repeat during the AI boom. In fact, dozens of well-known CSPs around the world are boldly entering the AI infrastructure race, challenging the established AI โpicks and shovelsโ companies (e.g., chipmakers and hyperscalers). Leading CSPs are realizing their unique advantages in domestic markets where local regulations over data residency and data sovereignty affect the use and distribution of AI (solutions), especially in highly regulated fields like government, healthcare, life sciences, financial services, energy, and utilities.
Many CSPs have already made major infrastructure investments to build up their data centers, GPU compute power, and network capacity to power the local AI economy in their core markets. Early signs of success โ judged by how fast they sell that capacity โ are fueling the momentum. This momentum will continue to build as long as supply-side shortages remain, evidenced by current customer waitlists and contracted revenue backlogs.
New Income Streamsย Andย Delivery Methods
Owning the underlying AI infrastructure is an asset to adjacent value realization, enabled by distribution of industry-specific AI solutions to the edge (i.e., AI as a service). CSPs are springing up AI factories โ often in partnership with chip manufacturers โ to provide AI model companies, cloud providers, and enterprise software companies with specific industry pedigree. Then, through contracts with original equipment manufacturers and end-user device manufacturers (e.g., personal electronics, smartphones, wearables), CSPs distribute those solutions to buyers and consumers on the edge (e.g., industrial IoT fleets, smart factories, businesses, consumers) by using their transport networks for two-way connectivity (via fiber-optic cables, 5G, DCI) and charging fees for each call.
To optimize and allocate the network capacity for the so-called โAI highwayโ and ensure that AI inference traffic doesnโt jeopardize the voice and data traffic and CX quality, CSPs are implementing multipurpose, convergent AI-RAN (AI-radio access networks) or O-RAN (open radio access networks). In the future, CSPs may transition to AI-native and AI-only networks and ultra-fast 6G.
The New Age Of AI-Powered Telcos Requires A Transformation
Besides the new and exciting commercial opportunities, CSPs are also changing from within by deploying AI in internal functions and operations. Along with the human workforce that must be upskilled and taught to work with AI, internal data foundations and fragmented systems must also be โupgradedโ and restitched. To unpack this and obtain examples of how CSPs are driving tangible outcomes, we interviewed dozens of telecom domain experts and current executives in charge of IT, AI, strategy, operations, transformation, and procurement. A view that most of them share is that โthe time for AI experimentation is over; the time for production has come.โ
Notable AI-led telecom transformation initiatives and outcomes include:
Network & Engineering
- Optimized energy consumption and hardware utilization
- Predictive fault detectionย (at cell level)ย and prevention
- AIโdriven network capacity planning
- Digital twins (replicas) of transport networks
- Autonomous,ย selfโhealing networks
Business & Operations Support Systems
- AI copilots for field technicians and service agents
- Agentic AI workflows across domains and functions
- Hyperpersonalizedย interactions across digital channels
- Proactive issueย detection andย resolutionย (pre-complaint)
- Enterprisewide agentic operating models to run, orchestrate, and govern AI agents
CSPs Are At A Pivotal Inflection Point โ Donโt Delay At The Start Of Your AI Voyage
AI is changing the economics of telecoms at a rapid pace. Those that successfully scale AI across their business will be best positioned to capture new growth and remain relevant in an increasingly intelligence-driven economy. Our latest report helps telecom digital, business, and technology leaders validate ideas, get inspired, and get started.
To helpย Forresterย clientsย on their AI voyage andย guide cross-functional impact,ย weโveย dedicated an entire content library housed in ourโฏAI research hub and collated over a thousand known business use cases for AI in Forresterโs AI Use Case Catalog.
Feel free toย get in touch withย meย orย scheduleย a guidance session by usingย this link.
Forrester clients can access the full report Original Post>
Enjoyed this article? Sign up for our newsletter to receive regular insights and stay connected.