AI checkpoint: The wins vs. the big rocks
AI Is a Telco Wake-Up Call — Not a Silver Bullet
At Mobile World Congress, one message came through clearly: AI isn’t a side project for telecom operators. It’s a checkpoint.
The industry is under pressure financially, structurally, and competitively. And while AI holds enormous promise, many operators aren’t structurally ready to capture its full value.
A recent Kearney report, European Telecoms 2026: In Need of a Health Boost, lays out the broader context. European telcos are grappling with squeezed returns, rising customer expectations, heavy capital requirements, and fragmented markets that make it difficult to achieve scale. You can read the full report here.
The question isn’t whether AI matters. It’s whether telcos can operationalize it.
The Reality: Ambition Is High. Readiness Is Low.
Across the board, there’s no shortage of AI ideas inside telecom organizations. From churn prediction to personalized offers to autonomous network optimization, the ambition is real.
What’s missing is alignment and infrastructure.
Operators are still running highly siloed organizations, often supported by dozens (or hundreds) of disconnected systems. In some cases, large acquisitions result in portfolios with zero overlapping platforms. Data is locked inside legacy OSS/BSS stacks. It’s not real-time. It’s not unified. And it’s rarely structured in a way that agentic AI systems can act on confidently.
The result? AI pilots everywhere. Scaled impact almost nowhere.
One executive put it bluntly: launching a new product can still take 10–12 months because of legacy constraints. If you can’t dynamically price or bundle services today, it’s hard to imagine delivering fully contextual, AI-driven offers tomorrow.
Where AI Is Gaining Traction
Despite the structural challenges, progress is happening, particularly on the efficiency side.
Most operators spend roughly 80% of revenue on operational expenses. Even small gains in cost-to-serve have an outsized impact on margins.
We’re seeing early deployments in:
- Call deflection and customer care automation
- Root cause analysis and anomaly detection
- Field force optimization and truck roll reduction
- Network operations and security convergence
These are pragmatic use cases. They don’t require a complete IT overhaul to start delivering value. And importantly, the ROI is measurable.
Revenue growth use cases like AI-driven acquisition, contextual cross-sell, or dynamic personalization are still maturing. The appetite is there, but most operators admit they aren’t yet structurally ready to execute at scale.
The Real Constraint: Data in Motion
Agentic AI changes the equation.
Unlike traditional automation, agentic systems need context. They require access to live data, clear relationships between systems, and the ability to reason across domains network, product, billing, customer history.
This isn’t about building another data lake.
It’s about enabling interoperability across legacy environments without ripping everything out. OSS and BSS platforms aren’t going anywhere. Many have decades of embedded logic and domain expertise. Replacing them wholesale isn’t realistic.
The more practical path forward is enabling data to flow securely and in real time, across systems. When data becomes accessible and structured, agents can act. When it stays locked in silos, AI stays stuck in demo mode.
Efficiency First. But Not Efficiency Only.
Most operators today are focused on efficiency gains, and for good reason. Moving OpEx even a few percentage points materially improves financial performance.
But there’s a second horizon.
AI may unlock business models that previously failed not because they were bad ideas, but because they were too expensive to run.
Security services. Home monitoring. Intelligent network-backed enterprise services. Context-aware bundles. Products that required heavy manual oversight in the past may become viable in an AI-assisted operating model.
The opportunity isn’t just to protect margins. It’s to revisit growth.
A Critical Moment for European Telcos
There’s also a regional dimension.
The U.S. tends to move quickly when capital markets see upside. Asia-Pacific often leads in deployment velocity. Europe operates in a more fragmented landscape, with nearly 90 operators competing across the region.
If AI becomes a structural advantage and markets begin rewarding AI-ready, data-ready operators with higher valuations — the gap could widen quickly.
The infrastructure buildout is already happening. Over a trillion dollars is being invested globally in AI-related infrastructure this year alone. The pressure to monetize that investment will only intensify.
So, Should Telcos Be Bullish?
Yes — but not complacent.
The last AI wave delivered chatbots and IVRs that never truly penetrated the core of the business. This wave is different. It’s capable of reshaping workflows, cost structures, and customer experiences.
But only if operators do the hard work beneath the surface:
- Break down data silos
- Expose systems through interoperable architectures
- Align around a clear AI “north star”
- Modernize without waiting for perfection
AI won’t fix structural fragmentation on its own. But it can amplify the operators that tackle it head-on.
For telecom, this isn’t hype. It’s a health check.
And for many, it’s time to start the treatment plan.
