How Telecom Companies Increase ARPU with AI Plan Recommendations

Updated on May 29, 2026

TL;DR

• Global mobile ARPU is on a structural downward path, and price hikes and cost cuts can’t fix it.
• Most telecom marketing efforts are still generic: the same plan is pushed to everyone, regardless of usage or lifecycle stage.
• AI-driven personalization changes this. It matches the right offer to the right subscriber at the right moment across chatbots, voice calls, WhatsApp, and more.
• Cross-selling with AI can increase ARPU by 5–20%; even a 1% reduction in churn saves millions annually.
• The two biggest barriers are fragmented subscriber data and earning customer trust. Solving the second one means leading with value instead of pushing the highest-margin plan

Every telecom website carries the same message: “Get our fastest plan. Unlimited data. Best value.” 

Big banners and high-budget OOG campaigns have created a culture of telecom marketing that wants to appeal to everyone. Consequently, the advertising also doesn’t fit anyone in particular. This generic approach to selling has been the telecom industry’s default for years. And it’s increasingly a problem, because ARPU (Average Revenue Per User) is under real structural pressure. 

This article decodes how AI-driven personalization in customer workflows can help ease this pressure. We’re going to cover:

1. The ARPU problem
2. The problem with generic recommendations
3. Where does AI change the equation?
4. The AI intervention playbook
5. Limitations of the AI-driven approach
6. Conclusion

The ARPU problem

According to PwC’s Global Telecom Outlook (2026), global mobile ARPU is expected to tick down to $6.20 in 2029 from $6.32 in 2024, with fixed broadband ARPU also under pressure. This is the result of consumers’ flat-to-down willingness to pay. PwC calls this a structural change, not a cyclical dip.

The on-the-ground view isn’t any more optimistic. At MWC 2025, a senior executive at Globe Telecom in the Philippines put it plainly: their ARPU growth had stagnated because of “increasingly homogeneous mobile data plans that don’t allow for differentiated network experiences, including premium offerings that command higher price points.” His diagnosis: “Consumers today demand more personalized and flexible offerings, and failing to meet these expectations can limit revenue potential.” 

Telecom executives around the globe have expressed similar sentiments. In a survey of 60 telecom CEOs and top executives globally conducted by McKinsey in February 2024, the most pressing concerns cited were:

• Profitability
• Competition from new business models,
• Regulatory constraints 

The $1.5 trillion industry has a major ARPU problem, and the usual responses to this crisis (price hikes, mergers, or cost cuts) have limited headroom. You can only squeeze margins so far before customers walk. And in most markets, they have somewhere to walk to.

The only way out is to market and sell better.

The problem with generic recommendations

The telecommunications industry is a  veritable treasure trove of data points:

1. Usage patterns
2. Device types
3. Roaming history
4. Data consumption trends
5. Top-up frequency 

Despite this, marketing campaigns from the industry have been remarkably stagnant and generic. As Sebastian Barros, an MD at Circles (a globally renowned vendor for telcos), says, “Telcos continue to sell gigabytes, speed, and 5G coverage, technical features that most customers neither understand nor truly value.”

A customer who uses 2 GB a month and calls their elderly mother twice a week is shown the same 100 GB unlimited plan as a remote worker streaming video calls eight hours a day. One of the users doesn’t need the extra bandwidth, and the other isn’t receiving the personalized campaign that will make them convert.

Traditional telecom marketing strategies often involve blanket promotions targeting large segments, many of which may not convert. 

This is wastage:

1. In marketing spend
2. In agent time
3. In missed upsell opportunities 

And predictably, the wastage compounds over millions of subscribers. In our experience, the solution to this is an approach that leverages the data telcos already have. That solution requires AI.

Where does AI change the equation?

AI-driven recommendation systems flip this logic. Instead of pushing the highest-margin plan to everyone, they analyze individual subscriber behavior and surface the offer most likely to be relevant, accepted, and retained.

1. AI-driven cross-selling can increase ARPU by 5–20%
2. AI-driven selling boosts B2C revenue by 2–4%
3. 31% of customers feel better understood by their provider when personalization is involved
4. 37% report less stress during their shopping experience.

These gains compound. A customer who gets a well-matched plan is less likely to churn, more likely to add services, and more likely to respond positively to the next outreach. But these techniques need a different marketing approach.

Where should you add personalized recommendations?

Most people think of AI recommendations as a website chatbot phenomenon: a chat widget that asks a few questions and suggests a plan. But the opportunity is broader:

Each of these is a moment where the right offer, delivered in the right way, can move ARPU. Since we serve enterprise telcos across the globe, we’ve also created a small playbook for such AI interventions. 

The AI intervention playbook

Not all personalization is equal. The interventions below tend to have the clearest commercial impact:

• Bandwidth nudges – When a subscriber consistently hits the ceiling of their current plan, that’s a natural, high-conversion moment to suggest an upgrade. The data already justifies it.
• Lifestyle-triggered conversations: a newdevice on the account, a family line added, or a relocation are all signals that a subscriber’s needs have changed. AI can catch these and open a relevant conversation.
• Bundle offers during support interactions – Customers who call in about a billing issue or network problem are already engaged. Surfacing a relevant add-on at that moment outperforms cold marketing campaigns by a wide margin.
• Proactive churn-risk outreach – Identifying at-risk subscribers before they decide to leave, then leading with a value-based offer rather than a generic retention discount. For a deeper look at how CX teams are doing this, read our guide on reducing telecom churn.
• Renewal-window upsells – Timed to contract milestones. A subscriber approaching the end of a contract will make a decision. Personalized upgrade offers at that moment are more effective.
• Post-complaint recovery sequences – After a negative service experience, an immediate upsell is tone-deaf. But a well-timed follow-up that acknowledges the issue and reintroduces value can recover the relationship.

For a detailed breakdown of how these journeys are structured end-to-end, our piece on AI customer journeys for telecom cross-sell and upsell covers the mechanics in depth. Additionally, we recommend using WhatsApp as a recommendation channel for this purpose.

Use WhatsApp as a recommendation channel.

One channel that often gets underestimated in the ARPU conversation is WhatsApp. The channel is mostly used for transactional alerts. 

1. Bill due
2. Payment confirmed
3. Outage update

But for a growing share of subscribers globally, it’s their primary channel for everything.

Telecoms leveraging analytics have seen 7% ARPU growth and 30–35% improvement in churn KPIs. WhatsApp automation, used for proactive plan recommendations in addition to support, can be one way carriers capture that upside.

Another method we recommend is through chatbot platforms. 

Using a Chatbot Platform

The quality of AI plan recommendations depends almost entirely on the platform carrying them. A basic FAQ bot can’t look at a subscriber’s usage history and generate a relevant upgrade pitch. A generative AI chatbot can.

One major provider using AI-driven personalization achieved a 50% boost in lead conversion rates for its B2B segment and added £80 million to their annual sales pipeline within just six months.

Choosing the right platform here is crucial. We’ve already evaluated the best AI chatbot platforms for telecom in terms of support capabilities, integration depth, and telecom-specific fit. 

Make the right offer, at the right moment.

The difference between a generic upsell and an AI-powered one lies in the timing and framing. Companies that excel in personalization generate 40% more revenue than average telecom providers. 

That gap comes from matching the right product to the right customer at the right moment, with messaging that makes the value obvious to that specific person.

As with all marketing, more specific is better. But there’s also a huge limitation to the AI-driven approach. 

Limitations of the AI-driven approach

AI-driven recommendations don’t start working automatically. Before a carrier can benefit from AI-driven recommendations, two foundational problems need to be solved: one of data fragmentation and one of customer trust.

1. The Data Fragmentation Problem

Most large telecom operators weren’t built for the kind of unified subscriber intelligence that personalization requires. They were built for network management, billing, and compliance, and the systems that handle each of these have multiplied and diverged.

A typical carrier today might have:

1. Subscriber usage data sitting in a network OSS
2. Billing history in a separate BSS
3. Device information in a CRM
4. Support tickets in a helpdesk tool
5. Payment behavior in a finance platform 

None of these systems  talk to each other in real time. For the AI to surface a relevant plan recommendation mid-conversation, it needs to pull from most or all of these simultaneously.

This is where the real implementation cost lives. Building the connectors between these systems requires ongoing integration work. The practical path forward involves:

• Prioritizing a customer data platform (CDP) or unified data layer that aggregates the signals most predictive of upgrade intent: usage patterns, plan ceiling hits, device age, tenure, and recent support contacts.
• Starting with a defined set of use cases rather than trying to integrate everything at once. A carrier that can reliably detect “subscriber consistently hits data cap” and trigger a contextual upgrade offer has already unlocked significant ARPU potential.
• Choosing AI platforms with pre-built telecom integrations rather than building from scratch. The integration burden is real, but much of it has already been solved by platforms purpose-built for the industry. This is one reason platform selection matters as much as the AI model itself.

The goal is a unified subscriber view that any customer-facing system can query in real time. Carriers that treat this as an infrastructure project rather than a commercial one tend to deprioritize it. Those who make personalization work tend to treat it as a revenue initiative from the start.

2. The Trust Problem

Even with perfect data, there is a ceiling on what AI recommendations can achieve if they feel like sales dressed up as service.

Customers have developed a finely tuned sense of the difference between a recommendation made for their benefit and one made for the company’s. The former builds the kind of trust that increases lifetime value. The latter erodes it in ways that don’t show up until churn does.

The discipline required here is commercial: optimize for the relationship, not the transaction. An AI that learns to surface the right offer at the right moment will outperform one that maximizes offer frequency over the long run.

That means building in guardrails: suppressing upsell prompts during active complaints, ensuring recommended plans genuinely fit the subscriber’s usage profile, and measuring outcomes over a 90-day window rather than a single interaction.

Conclusion

Telecom ARPU pressure is structural. Carriers can’t wait for the market to improve or hope that a 5G rollout magically increases willingness to pay. The opportunity is in selling smarter using data they already have to match subscribers with plans that actually fit their lives.

The carriers that treat AI-driven personalization as a core commercial strategy will likely find it’s one of the few ARPU levers that doesn’t require launching a new product or raising prices across the board.

If you want to implement these features across the customer journey, feel free to book a call.