Agentic AI is quickly moving from experimentation to implementation across the telecom industry. Operators are exploring how autonomous AI systems can improve customer service, predict churn, personalize offers, streamline operations, and unlock new revenue opportunities. And for many, these technologies are already becoming embedded in operations. IDC predicts that 40% of Global 2000 job roles will involve working alongside AI agents by the end of this year.

One challenge stands out: AI agents are only as effective as the data they work from. And for many operators, the subscriber data underpinning those ambitions remains fragmented, inconsistent, and spread across years of disconnected systems.

If operators deploy agentic AI before addressing those foundations, they risk making the wrong decisions at scale.

The hidden cost of telco-techco transformation

Many operators created this challenge while pursuing entirely sensible business goals. For more than a decade, telecom providers have been evolving beyond connectivity. Faced with commoditized network services and mounting competitive pressure, they’ve expanded into streaming, gaming, digital lifestyle services, financial products, and broader partner ecosystems.

The strategy was always sound. Connectivity alone is no longer enough to drive meaningful growth. The problem is that each new service often arrives with its own billing engine, entitlement system, customer database, payment workflow, and partner integration. Large-scale business support system (BSS) transformation programs frequently stretched over several years, while customer expectations and market dynamics evolved far more quickly. In many cases, by the time new systems were deployed, operators were already introducing new integrations and workarounds to support evolving requirements. It's also common for the executives who championed these programs to have moved on before the transformation reaches its objectives.

As a result, operators now manage highly complex subscriber environments where valuable customer information exists across multiple systems, making it difficult to establish a consistent, real-time view of the subscriber. This becomes more problematic as operators expand their role as digital service aggregators. Streaming partnerships, gaming bundles, financial services, and connected lifestyle offerings all generate new customer interactions, entitlements, and revenue relationships. The opportunity isn't constrained by access to services. It's constrained by the commercial models underpinning them – and how effectively those services are monetized and managed.

Why agentic AI amplifies data problems

The next phase of AI differs fundamentally from earlier automation initiatives. Traditional ML/AI models helped organizations analyze data and identify opportunities. Agentic AI takes the next step by acting autonomously. From a customer management perspective, that might mean detecting churn risk, creating retention offers, adjusting pricing, triggering campaigns, and orchestrating customer journeys without waiting for human intervention.

These types of concepts come with potential pitfalls. An AI agent attempting to retain a customer needs access to accurate billing history, subscription status, entitlement data, payment information, engagement signals, and partner service activity. If those systems fail to communicate clearly with one another, the agent's recommendations become less reliable.

People often reference "garbage in, garbage out" when discussing AI. With agentic AI, the greater concern is that poor data can drive confidently executed but fundamentally wrong decisions – at speed and at scale.

Consider a simple retention scenario. An AI agent identifies a subscriber as a high churn risk and automatically generates a personalized offer. But if the underlying data doesn't accurately reflect the customer’s lifestyle habits, content preferences, or mobile usage, the intervention may be at best irrelevant – or worse, another clunky customer interaction that lowers satisfaction.

Many legacy and highly customized BSS environments weren’t originally designed to provide a unified view of subscriber behavior across multiple services, partners, and touchpoints. As operators expand into streaming, gaming, digital lifestyle services and partner ecosystems, understanding the full subscriber relationship becomes increasingly difficult.

At the same time, many general-purpose AI tools lack the industry-specific context required to understand telecom subscription businesses. Subscriber entitlements, bundle structures, partner revenue-sharing arrangements, billing relationships, and regulatory obligations are highly specialized domains. Without clean, connected subscriber data and clear commercial logic, even sophisticated AI systems can struggle to generate recommendations that align with business realities.

The goal should be ensuring new AI agents have the context, intelligence, and operational visibility needed to act effectively.

Entering the intelligence economy

The wider subscription industry is entering what can be described as the intelligence economy. For decades, competitive advantage came from building networks. Later, it came from delivering digital experiences and content. Today, advantage increasingly comes from understanding subscriber behavior in real time and translating that understanding into action.

The operators that succeed in this next phase will be those that can unify subscriber identities, consolidate entitlement logic, connect partner ecosystems, and establish trusted data foundations across their commercial operations. Only then can agentic systems deliver on their promise.

This becomes especially important as operators pursue more sophisticated growth strategies. Every new partner relationship creates additional data and customer touchpoints. The challenge has moved beyond subscriber acquisition and service expansion. It’s now about understanding customers across increasingly complex ecosystems and responding intelligently to changing needs.

Agentic AI can absolutely help operators launch offers faster, personalize experiences more effectively and improve retention. But none of that happens without clean, connected subscriber data. The biggest challenge here isn't designing smarter agents – it's fixing the subscription foundations those agents depend on.