With tens of millions of subscribers across multiple product lines like global system for mobile communications, broadband, WiFi and dongles, telcos must also manage multiple accounts-including prepaid or postpaid-bundled under one bill. This doesn’t even factor in the intricate customer journey that encompasses lead capture, onboarding, nurturing, upselling and cross-selling, loyalty, win-back and, increasingly, data monetization through curated audience segments.
While ML can help manage many of these complexities for marketing teams, it’s important to be clear: ML is not the starting point.
The journey begins with the customer-and just as critically, with the investor. Telcos must offer the best network quality, voice, data and sensible plans. Investors expect the business to stay profitable while keeping customers happy. Both are necessary
Some telcos are obsessed with distancing themselves from the “old-fashioned” label of being a telco. This leads to digital rebranding efforts that try to paint a picture of something beyond traditional telecom-but without a clear direction.
Yes, there is room to grow beyond the traditional telco, but there is still a lot to gain by consolidating revenue, improving loyalty and charging a premium for quality services. Before fully embracing ML, telcos need to do their homework. They must critically assess their current state and accept:
ML in customer value management has come a long way-from simple models like preferred channel to offer prioritization and next best offer (NBO). The latest wave of Al agents is pushing marketing autonomy further, where businesses set objectives and ML handles execution details.
Zero-touch campaigning is the ultimate goal, a strategy that automates the entire campaign lifecycle. Achieving this involves progressing from current NBO-level maturity, through Al agent-enabled semi-autonomy, to a future where businesses provide goals, budgets and boundaries, and ML optimizes end-to-end.
But no business can operate solely at that high level of abstraction. Guardrails and scaffolding are essential. A strong data platform is the foundation. Telcos need identity stitching, engagement stitching, micro-journey curation and a unified customer view. ML supports these building blocks by removing noise, uncovering correlations and delivering predictive analytics. The focus should be on the function, not on marketing labels like customer data platform.
After that, the jump to zero-touch campaigning is steep. ML moves from data points to the “people” and “business” plane-where emotions, perceptions, money and timing matter. This stage depends on constrained optimization, reinforcement learning, prescriptive analytics and Al agents.
To develop these abilities, marketing teams must move beyond the linear funnel and take a non-linear route.
Likewise, marketers at telcos should focus on their transformation journey, not just tools or vendors. They need to spot areas ready for autonomy and gradually hand those over to Al agents. Good examples of areas ripe for automation include frontline personalization, win–back campaigns, churn arrest and strategic ambitions in specific markets or products
To better understand where telcos stand and where they are headed, it helps to think about ML maturity as three levels:
Level O: Predictive And Diagnostic Analytics
At this stage, ML works primarily with descriptive and predictive models-removing noise, finding correlations and analyzing data to generate insights. These models mostly operate on raw data and focus on answering questions like “What happened?” or “What might happen?” This is the foundation, where data quality and integration are critical.
Level 1: Localized Prescriptions And Reinforcement Learning
Here, ML starts making prescriptive decisions about what actions to take to influence customer experience (ex) or business outcomes (like revenue). This level typically focuses on one dimension at a time–for example, optimizing offers to reduce churn or increase sales. Models incorporate feedback loops and learn over time using reinforcement learning techniques, but operate in isolated functions or channels.
Level 2: Al Agents With Greater Autonomy And Fusion Of ex And Dollars
The most advanced stage sees Al agents that manage multiple objectives simultaneously-balancing customer experience with business profitability. These agents operate with significant autonomy, making strategic decisions while respecting guardrails set by the business. This stage represents the full potential of zero-touch campaigning, where goals and budgets are set upfront, and ML-driven agents execute campaigns end-to-end with minimal human intervention.
This progression helps telcos realistically chart their journey, identify gaps and set achievable milestones. Understanding where you are on this spectrum is more valuable than rushing to adopt the latest shiny tech without context.
Often, the best transformation starts with silent introspection. Having worked with over 40 telcos globally, I have seen many in a state of continuous customer value management transformation-sometimes driven by their own will, sometimes by peer pressure.
This article is meant as a subtle mirror. It doesn’t push a particular solution. It invites telcos to reflect honestly on where they stand, what they want and how they want to move forward. Like deciding how to travel from London to Paris-by land, rail or air-there’s no one right answer. Each path has different merits, costs and risks
ML is a lever-not the driver. The true driver is a clear business vision balancing customer satisfaction and investor returns. When telcos approach ML as a tool in their hands rather than a magic wand, and plan their journey thoughtfully, they stand to unlock real value.
