AI Lessons from Cell C’s 2025 Digital Turnaround

Cell C didn’t “win 2025” by spending like a bigger rival. It did it by getting smarter about how it builds networks, how it serves customers, and how it partners. That’s why its 2025 story matters to anyone building AI-powered digital services in South Africa—especially in e-commerce, where margins are tight and customer expectations are brutal.

The headline event was Cell C’s return to public markets with a JSE listing (ticker: CCD) after a disciplined turnaround. But the more interesting signal is what sat underneath that milestone: cloud-native infrastructure, digital-first customer journeys, and distribution partnerships designed for scale. If you run an online store, a marketplace, a subscription service, a fintech app, or any “digital-first” business, there are practical lessons here.

This post is part of our series on How AI Is Powering E-commerce and Digital Services in South Africa, and I’m going to take a stance: the companies that win in 2026 won’t be the ones with the flashiest AI demos. They’ll be the ones that treat AI as an operating model—built on cloud foundations, clean data, and customer-centric automation.

The real shift: from owning everything to orchestrating partners

Cell C’s core move is simple: a capex-light, partnership-first approach that prioritises outcomes over ownership. In telecoms, that means relying on network partnerships while focusing internal effort on customer experience, product design, and service quality.

For e-commerce and digital services, the parallel is obvious. Most South African businesses can’t afford to “own the stack” end-to-end—warehouse automation, custom apps, in-house data science, 24/7 support, fraud tooling, and personalised marketing. The better model is to orchestrate: choose strong platforms, integrate cleanly, and differentiate where it counts.

Here’s the AI angle: modern AI systems are only as good as the plumbing beneath them.

What a “capex-light” model looks like in digital services

Translate Cell C’s approach into a digital commerce context and you get a playbook like this:

• Use cloud infrastructure (for elasticity and cost control) instead of buying hardware “just in case.”
• Build modular services (payments, messaging, CRM, analytics) rather than one monolithic system.
• Treat partnerships as product features: delivery networks, device ecosystems, retail distribution, and customer support channels.

The businesses that pull this off don’t just save money—they move faster. And speed matters when your competitors can copy your product in weeks.

Cloud-native infrastructure is a prerequisite for AI-driven customer experience

Cell C highlighted live VoLTE testing on a fully virtualised, cloud-native IMS platform hosted on AWS, positioning next-generation voice services as something you can deliver with software-centric architecture.

If you’re in e-commerce, you might think, “Cool telecom story, not my world.” It is your world. Because the same foundation—cloud-native, virtualised services—is what makes AI practical at scale.

Why cloud-native matters for AI in South Africa’s digital economy

AI workloads are spiky. Your traffic spikes on payday weekends, Black Friday, Christmas promotions, and back-to-school. (And yes, December 2025 is exactly when you feel this.) Cloud-native architecture lets you scale up inference, search, recommendations, and fraud checks when you need them—without permanently carrying that cost.

Three concrete ways cloud-native design supports AI-powered e-commerce:

1. Real-time personalisation: product recommendations, next-best-offer, and personalised bundles require fast data pipelines.
2. Automated customer support: chat and voice assistants need reliable APIs into orders, returns, account status, and delivery tracking.
3. Operational automation: demand forecasting, replenishment suggestions, and delivery ETA prediction improve with more data and faster processing.

A useful one-liner to keep on your wall:

If your systems can’t respond in real time, your AI will only ever be a reporting tool.

The Cell C App story: digital-first doesn’t mean “pretty UI”

Cell C redesigned and relaunched its app to make it easier for customers to manage usage, buy value, and access support quickly. That’s not just a UX refresh. It’s a shift toward self-service as the default, with support and sales embedded in the same journey.

E-commerce businesses often copy the surface layer (new app screens, new onboarding, new checkout). The better move is to copy the underlying principle: remove friction and let the customer complete the job without asking for help.

Where AI fits in app-led customer engagement

If you’re building digital services, AI should do three jobs inside your app:

• Predict intent: why is the user here right now—browse, compare, buy, return, complain, or track?
• Shorten time-to-resolution: route them to the right answer, agent, or workflow immediately.
• Increase confidence: show shipping reliability, sizing guidance, product fit, stock accuracy, and transparent pricing.

Practical examples (that don’t require a moonshot AI budget):

• A returns assistant that asks two questions, auto-generates the return label, and books a pickup.
• A “help me choose” flow that learns from clicks and purchases and updates recommendations instantly.
• A support triage classifier that tags tickets by urgency and topic so agents don’t waste the first 5 minutes.

Cell C also introduced “Try Before You Buy” across app, online, and in-store touchpoints. The deeper lesson isn’t the offer itself—it’s the omnichannel consistency. AI performs better when your journeys are consistent because the data is comparable and the outcomes are measurable.

Distribution partnerships: the underrated AI advantage is better data

Cell C leaned into distribution partnerships to expand access to SIM cards, devices, and offers in high-traffic retail locations. That’s a classic growth lever, but it also has a modern implication: distribution creates more customer touchpoints, and touchpoints create more behavioural data.

For e-commerce, distribution can mean pickup points, retail partners, marketplace listings, or device ecosystems. More channels can be messy, but if you design it properly, it becomes fuel for AI.

Turning channel sprawl into an AI-ready system

Most companies get this wrong by bolting on channels without standardising data. The fix is boring, but it works:

• Use one customer identity approach across channels (even if it’s probabilistic, not perfect).
• Standardise event tracking (view, add-to-cart, checkout, delivery status, return initiated, refund completed).
• Centralise product and pricing rules so promotions don’t drift by channel.

Once you have that, AI can deliver measurable improvements:

• More accurate demand forecasting because you’re not blind to offline or partner-driven demand.
• Lower fraud and chargebacks because patterns show up across channels.
• Better lifetime value predictions because you see the full customer story.

Enterprise cloud partnerships: what SMEs should copy (and what they shouldn’t)

Cell C signed an MoU with Huawei Cloud around cloud services and digital infrastructure, positioning itself to support enterprise and public-sector digitalisation.

SMEs and mid-market e-commerce brands should take two things from this:

Copy this: choose partners that make you faster

Pick partners that reduce time-to-market for:

• customer communications (messaging, email, in-app notifications)
• analytics and experimentation (A/B testing that’s actually trustworthy)
• support operations (ticketing, knowledge base, automation)
• security and compliance basics

Speed isn’t about “shipping features.” It’s about shipping learning. You want fast feedback loops so your AI models and automation rules get better weekly, not yearly.

Don’t copy this: vendor lock-in without an exit plan

Partnership-first is smart. Permanent dependency isn’t.

If AI is central to your customer experience, ensure:

• your data can be exported in a usable format
• your core customer events aren’t trapped in a proprietary tool
• you can swap components (CDP, search, support bot) without rewriting your entire product

That’s not paranoia. It’s basic resilience.

People also ask: what does this mean for AI-powered e-commerce in 2026?

Will telecom improvements actually affect online retail?

Yes—directly. Better connectivity improves checkout completion, reduces payment failures, supports richer product content (video, AR, live chat), and makes app-based commerce smoother in more geographies.

Is “digital-first” just a cost-cutting play?

No. Done well, it’s a service quality play. Self-service reduces wait times, but the real win is consistency: the customer gets the same outcome whether they use app, web, or store.

What’s the most practical AI investment for a South African online store?

Start with support automation and product discovery. They touch revenue and cost immediately:

• deflect repetitive tickets (order status, returns policy, delivery updates)
• improve search and recommendations so customers find products faster

What to do next: a simple 30-day plan you can actually run

If Cell C’s 2025 story tells us anything, it’s that momentum comes from execution, not announcements. If you’re trying to build AI-powered customer engagement in e-commerce or digital services, here’s a tight plan for the next month:

1. Audit your customer journeys: list the top 10 tasks customers try to complete (buy, track, return, cancel, change address, get support).
2. Instrument the basics: ensure you capture events reliably across app/web/support.
3. Pick one automation wedge: order-status assistant, returns flow, or support triage.
4. Measure two numbers: resolution time and conversion (or deflection rate). If you can’t measure, you can’t improve.
5. Tighten the loop weekly: update FAQs, automation rules, and model prompts based on real transcripts and outcomes.

The stance I’ll leave you with: AI in South African e-commerce is becoming an expectation, not a differentiator. The differentiator is whether your infrastructure, data, and operating rhythm can keep improving after you ship.

Cell C’s turning point in 2025 wasn’t luck. It was a decision to rebuild around digital-first execution and scalable partnerships. If your business wants more leads and better retention in 2026, the question isn’t whether you “should do AI.” It’s whether your customer experience is structured in a way that AI can actually help.