At this year's CANTO conference, one slide stopped the room. Ranking the world's regions across the skills that will define the next decade, it laid out an uncomfortable truth: across much of the Caribbean - as well as several other regions - the workforce still scores low on the technical skills of the moment, with AI and big data right at the bottom of the list.
To a lot of leaders, AI can feel like a fast-moving wave you're already behind on - expensive, complicated, and risky to get wrong. If your teams aren't up to speed technically, the instinct is often to delay investment – or forgo it altogether.
But being light-on-skills doesn’t reflect what your people can do; it's simply a sign they haven't been given the tools yet. And the payoff is huge: AI is already helping telecom teams sell more, serve customers faster, and claw back hours of everyday work.
In this edition of Telco Talk, we'll examine how transformative AI can be for telcos today, as well as lay out a practical, low-risk roadmap for successfully introducing it to your organization.
Where The Opportunity Is: AI Worth Investing In
AI has moved well past the pilot stage, but the term still means very different things to different people. The clearest way to think about it is in two parts: the foundational investments you make - an AI agent, and a connection to your commerce platform - and the platform-native features you should look for in your commerce and point-of-sale systems, which deliver value on their own regardless of whether you've adopted an agent. Here's what each is, in plain terms, and what it returns.
Investment 1: An AI Agent For Everyday Work
The first and most important step isn't a niche tool - it's adopting a capable AI agent your whole organization can use day to day. It's worth being clear on what that means, because it's different from the chatbots most people have tried.
A tool like ChatGPT is a general-purpose assistant: it's knowledgeable, but it only knows what it was trained on and whatever you paste into the conversation. An AI agent is different in two important ways. First, it can carry out tasks rather than just answer - working through multiple steps and taking actions on your behalf. Second, and more importantly, it connects to your organization's ecosystem - SharePoint, your knowledge base, internal documents, ticketing systems, and more - so its answers are grounded in your actual business context rather than generic internet knowledge. (It can still run everyday chatbot-style conversations too; that's just the surface.) The real power is that when someone asks a question, the agent can draw on what your company actually knows. Realm is one example of this kind of tool - and is what we use at Maplewave.
On its own, an agent already reclaims enormous time across every department:
- Drafting & Documentation: Producing first drafts of documents, proposals, reports, and customer communications in minutes, which a human then edits and approves - turning the hardest part (the blank page) into a starting point.
- Technical & IT work: Speeding up the everyday technical tasks a telco runs on - writing and testing integration scripts, troubleshooting configurations, documenting systems, and (for teams that build or customize their own software) generating and reviewing code - whether that work happens in-house or alongside a vendor.
- Analysis & Research: Summarizing long documents, pulling scattered data into a manager-ready briefing, and turning raw operating numbers into the questions worth asking in the next meeting.
- Everyday Operations: Automating repetitive back-office steps, routing requests to the right place, and preparing routine internal communications.
Benefit: This is where an agent stops being a personal productivity trick and becomes an organizational advantage. When every department is using the same system, the gains compound.
Technical teams resolve issues and document systems faster; marketing and sales turn around campaigns and proposals in a fraction of the time; operations and finance get answers from their own data without waiting on an analyst; leadership gets briefings and summaries on demand. Work that took days takes hours, and specialists can now spend their time on judgment instead of grunt work.
Just as valuable is the everyday fluency it builds. The confidence and habits that everything else in this roadmap depends on. The organizations pulling ahead aren't just letting departments haphazardly try out different tools and calling it an “AI strategy”; they've put a single agent to work across every team, so that using AI becomes simply how work gets done rather than a novelty that a handful of specialists touch. That, more than any single feature, is what "AI adoption" really means: not a tool that got installed, but a habit that took hold. It's also the lowest-risk place to start - the tasks are internal, the stakes are low, and the payoff is immediate.
Investment 2: Connect The Agent To Your Commerce Platform (Via MCP)
Once your team is comfortable with an agent, the next investment multiplies its value: connecting it to your business platform. Model Context Protocol (MCP) is essentially a standardized bridge that links your AI agent to your platform's data and functionality - securely, and through the AI client you've already adopted.
Suddenly the agent isn't just drafting documents; it can read your real business data and take action inside your retail systems.
Your team starts working in plain language:
- Create a purchase order.
- How did the north region perform last quarter?
- Build a personalized offer for this segment.
Benefit: The agent graduates from a general-purpose assistant to one that runs on your operations. Tasks that once meant hunting through screens or waiting on a report become a question anyone can ask - spanning routine operations, live reporting, and customer-facing work all at once.
It's also where personalized / next-best offers come to life: because the agent can now reach each customer's usage, billing history, and context, it can generate the single most relevant offer for them on demand, rather than blasting the same promotion to everyone. The payoff there is real: operators today push 15–40 messages per customer per month at response rates below 1%, while genuine personalization has delivered ARPU gains of 5–12% and a better customer relationship rather than spam.
Platform-Native Features To Look For
Separate from the agent and MCP investments above, there's a set of high-value AI capabilities some commerce platform providers are beginning to offer. These are built-in features that are worth pressing your providers on (or making a switch altogether).
AI-Suggested Replenishment
- Instead of managing inventory based on a manager's gut feeling, the platform watches current stock, how fast each item is selling, and demand patterns (seasonality, promotions, local trends), then generates a specific restock/transfer recommendation for staff to review and approve.
- Benefit: Fewer stockouts and lost sales, less capital tied up in dead inventory, and hours of counting replaced by a decision that takes minutes - one of the most direct routes to protecting margin.
AI-Assisted Upselling
- No associate can hold every plan, add-on, and promotion in their head. Built into the point of sale, an upselling assistant analyzes what the customer already has and what similar customers value, then surfaces the most relevant add-on or upgrade in the moment.
- Benefit: Higher attach rates and bigger baskets without pressure selling; AI-assisted recommendations have driven upsell-conversion improvements of 20–45% for adopting operators, and they double as on-the-job coaching for newer staff.
The One Thing That Decides The Return
Here's the catch that can get overlooked: every one of these tools needs a human to decide whether to trust the output, then refine it, and act on it. That's not a limitation to work around; it's where the value is actually captured. The industry consensus is that AI should enhance, not replace, human judgment, which is precisely why an adaptable, curious workforce is the asset that makes the whole investment pay off.
And it's why exposure - the reason to lead with an agent - matters more than any single purchase. Even in mature markets, 75% of retail executives call AI a top strategic priority, yet only 16.5% can quantify a return on it. The tools get bought; the return depends on whether people actually use them well. For an organization with an adaptable workforce and no legacy AI habits to unlearn, that gap between spend and return isn't a warning, it's the opening. Close it, and the same tools that sit idle elsewhere become a genuine advantage.
Roadmap: How To Actually Incorporate AI
Knowing the opportunities is one thing; turning them into results is another. Here's a practical sequence any organization can follow.
1. Build AI Literacy & Familiarity First
Before any project, get people using AI. Broad, hands-on exposure to everyday AI tools for drafting, summarizing, answering questions builds the fluency and confidence everything else depends on. This is the cheapest, fastest step and the one most often skipped. Start here and start wide: the goal is for AI to feel normal, not novel.
2. Choose The Projects With The Most Benefit
You can't do everything at once, so pick deliberately. Weigh each opportunity by impact versus effort and risk: internal-efficiency wins (drafting, testing, analysis) are usually the quickest returns, while customer-facing tools like upselling and personalized offers carry bigger revenue upside but need more data and trust. Sequence from low-risk, high-visibility wins toward the higher-value plays.
3. Choose An AI Engine For Everyday Work
Standardize on a capable, secure AI platform your whole organization can use day to day, rather than a scatter of one-off tools. A single trusted engine for chat, drafting, analysis, and agents makes adoption easier to support, govern, and measure. It’s also what turns literacy into a durable habit across departments.
4. Make Sure Your Platform Providers Can Deliver The AI Features You Need
Many of the highest-value capabilities - suggested ordering, personalized offers, conversational data access - depend on your underlying commerce and operations platforms exposing them. Press your providers now: can they surface AI-driven recommendations, connect to an AI layer, deliver the features on your roadmap? The best workforce in the world can't use capabilities the platform doesn't offer. Choose partners who are building toward where you're going.
5. Don't Wait - The Gap Compounds
AI capability builds on itself: the fluency, data, and habits you establish now make every future step faster, while organizations that wait fall further behind each quarter. First movers in a market don't just adopt sooner; they set the standard customers come to expect. The best time to start was yesterday. The second-best time is now.
Final Thoughts
Starting with AI can feel daunting, but the path is simpler than it looks. No matter where you’re starting, build familiarity first, pick the projects with the most benefit, standardize on an engine your whole team can use, make sure your platforms can deliver the features you'll want next - and, above all, start now. The gap only compounds, and the organizations that move first won't just adopt faster; they'll set the standard everyone else has to catch up to.
