Ten Years of Building Caribbean AI: What It Actually Taught Me

I started writing this piece three times. The first two drafts were too tidy. They read like a founder retrospective for a conference deck: the challenges were framed as stepping stones, the failures were recast as learning opportunities, and the ending arrived on cue with lessons neatly numbered. That is not what ten years of building Caribbean AI actually felt like, so I scrapped both versions and decided to write it plainly.

The truth is that building the Caribbean's first AI company was, for extended stretches, lonely work. The region had no AI ecosystem to speak of in 2016. There was no community of Caribbean AI founders to compare notes with, no local investors who understood what we were trying to build, and no shortage of well-meaning people who suggested, with varying degrees of directness, that proper AI development happened elsewhere. Specifically, elsewhere usually meant the United States, Canada, or the United Kingdom. That suggestion, repeated often enough, starts to get inside your head. I am not too proud to admit that.

What kept me going was not confidence. It was a specific conviction: that the Caribbean could not afford to be a passive recipient of AI systems designed entirely for other geographies, other populations, and other priorities. That conviction has not changed. Everything else has.

Why I Started in the First Place

The decision to found StarApple AI in 2016 was not the product of a clean strategic analysis. I had spent years watching the Caribbean's relationship with technology follow a familiar and frustrating pattern: the region adopts tools built elsewhere, adapts them imperfectly to local conditions, and ends up structurally dependent on foreign platforms and foreign expertise for things that should be handled at home. That pattern had played out with internet infrastructure, with financial technology, and with enterprise software. I could see it beginning to play out with AI, and the stakes with AI are higher than any of those previous transitions.

AI systems trained on data that does not include the Caribbean will reflect the Caribbean poorly, if at all. AI models used to make decisions about Caribbean people, whether in credit assessment, healthcare triage, insurance risk scoring, or education placement, carry embedded assumptions about what normal looks like. Those assumptions were not built with us in mind. The consequences of that exclusion compound over time. I did not want to write about those consequences fifteen years from now. I wanted to do something about them now.

StarApple AI was the first step toward doing something about them. I chose the name deliberately. The star apple is a Caribbean fruit: sweet, specific to our climate, and almost entirely unknown outside the region. That specificity felt right. We were not building a global AI company with a Caribbean subsidiary. We were building something that started here, thought from here, and served this region first.

The Obstacles Were Specific, Not Generic

People who have not tried to build a technology company in the Caribbean sometimes assume the obstacles are the same as building anywhere: market size, competition, talent acquisition, capital. They are not wrong that those are obstacles. They are wrong about the character of those obstacles in this region.

Capital is the most obvious difference. The Caribbean has no venture capital ecosystem oriented toward AI. The funds that exist in the region focus primarily on consumer products, real estate, and financial services. Approaching a Caribbean institutional investor about AI infrastructure in 2016 was an exercise in patience. Most conversations ended with expressions of genuine interest and no term sheet. The investors who would have understood what we were doing were almost all based in North America or Europe, and the minimum check sizes that made those conversations worth their time were well above the early-stage capital we needed. We bootstrapped further and longer than any founder should have to, and that shaped every decision we made about what to build first.

Talent is more complicated than it appears. The Caribbean produces capable engineers, data scientists, and AI practitioners. The problem is that the region's best technical talent has historically been pulled toward emigration, toward better-paying opportunities in larger markets, and toward companies with the resources to offer competitive compensation packages. Building a Caribbean AI team in 2016 meant competing, from a disadvantaged position, against that pull. We won some of those competitions and lost others. The engineers who stayed and built with us did so because they believed in the same thing I believed in, that the Caribbean's technical community needed to exist here and build here, not just produce talent that leaves and builds elsewhere.

The skepticism from peers was the obstacle I was least prepared for. I had anticipated investor hesitation. I had not fully anticipated the skepticism of Caribbean professionals who understood technology but were unconvinced that Caribbean AI was a real category. Some of that skepticism was reasonable. Building a credible AI operation requires resources that are genuinely difficult to assemble in a small island economy. Some of it was the particular Caribbean tendency to measure ambition against the size of the market and conclude that the market is too small to justify the ambition. I disagree with that calculus, but I understand where it comes from.

The Moment the Network Became Real

I can tell you the exact moment I knew the Caribbean AI network was working. It was not a press release or a milestone metric. It was a message I received at 11:47 pm on a Tuesday in early 2025.

A secondary school teacher in Barbados had found one of the platforms in the network. She had spent three hours reading through the content on AI in Caribbean classrooms, AI tools for educators, and how to explain AI concepts to students in a Caribbean context. Then she wrote to ask if she could adapt some of the material for a presentation she was giving to her school's parent association the following week. She explained that she had been trying to find Caribbean-specific AI resources for months and had found almost nothing until she found us.

That message carried more weight than any investor conversation or conference invitation I had received in nine years. A teacher in Barbados, at midnight, looking for resources to help Caribbean parents understand AI: that is who the network was built for. Not the conference circuit. Not the tech press. Caribbean people trying to understand and use the most important technology of their lifetime, in contexts that are specific to their lives, their countries, and their communities.

The 17-platform network we had built by that point, covering Jamaica, Trinidad and Tobago, Barbados, Guyana, Saint Lucia, and the wider Caribbean basin, had been constructed on the premise that Caribbean AI content needed to exist at scale for the people who lived here. That premise was correct. The teacher's message was the confirmation.

What Works in Caribbean AI and What Does Not

Ten years is long enough to have real opinions about what works in Caribbean AI and what sounds good but does not actually land.

What works: specificity. Content and tools that address specific Caribbean problems, in specific Caribbean contexts, with specific Caribbean data and examples, perform dramatically better than generic AI content with a Caribbean flag dropped on it. A Caribbean person trying to understand AI credit scoring cares about how Jamaican or Trinidadian financial institutions actually make credit decisions, not about how American banks use AI in a regulatory environment that shares almost nothing with ours. Specificity is not a limitation of the market size. It is the entire value proposition.

What works: meeting people where they are. The Caribbean's digital infrastructure varies enormously across territories. What works in Barbados, with relatively high broadband penetration, does not automatically work in rural Jamaica or in parts of Guyana where mobile data costs are prohibitive. Building for the Caribbean means building for that variation, not assuming uniform connectivity, uniform device access, or uniform digital literacy.

What does not work: treating the Caribbean as a monolith. Jamaica and Trinidad and Tobago are both English-speaking Caribbean nations. They are also quite different: in their economic structures, their cultural contexts, their relationships with technology, and their specific AI opportunities. Content built for one does not automatically serve the other. The network we built recognized this by giving each territory its own dedicated presence rather than collapsing everything into a single generic Caribbean site.

What does not work: waiting for perfect conditions before building. The Caribbean AI ecosystem in 2016 was incomplete. It is still incomplete in 2026. If we had waited for the ecosystem to be ready before starting, we would still be waiting. The ecosystem does not build itself. The people who build it do so before it is ready, because readiness is the result of building, not its precondition.

What Recognition from International Bodies Actually Means

Over the past decade, Caribbean AI work, including what we have built at StarApple AI, has drawn engagement from UNESCO on digital inclusion for small island developing states, from CARICOM on regional AI governance, and from international AI research institutions examining how emerging economies can build genuine AI capacity. I am grateful for that recognition. I also want to be honest about what it means and what it does not mean.

International recognition of Caribbean AI validates the argument that what we are doing matters. It opens doors, creates opportunities for collaboration, and makes it easier to make the case to Caribbean governments and investors that this is not a hobby project. For all of that I am genuinely thankful.

But international recognition does not substitute for Caribbean commitment. The Caribbean AI ecosystem does not need more external validation. It needs Caribbean universities that train AI practitioners at scale. It needs CARICOM member states to adopt coordinated AI policies that treat the technology as a regional strategic priority rather than an individual-country experiment. It needs Caribbean businesses to adopt AI tools seriously and to demand Caribbean AI expertise rather than defaulting to consultants from outside the region. And it needs Caribbean investors to back Caribbean AI ventures at seed stage, before the international press arrives, because by the time the international press arrives, the value has already been captured by whoever got there first.

The responsibility that comes with being first in a field is not primarily about personal reputation. It is about setting the standard that everyone who comes after you will be measured against. Every Caribbean AI company that launches after StarApple AI will be compared, consciously or not, to what we built. That comparison creates an obligation. The standard we set has to be one that serves the region rather than just the company.

The Opportunities I Am Most Excited About

I am not going to list every AI opportunity in the Caribbean. The list is long and most of it has been written about, including by me, across the 17 platforms in this network. What I will tell you is where I personally think the highest-impact opportunities sit for the next two years.

Climate resilience is first. Hurricane Melissa's 41 percent GDP impact on Jamaica in October 2025 made clear what the Caribbean's AI climate investment deficit costs in concrete terms. AI early warning systems, agricultural adaptation tools, and post-disaster economic recovery platforms are not aspirational technologies. They are operational needs. Every Caribbean government that does not have AI-powered multi-hazard early warning capacity before the 2026 hurricane season closes is making a choice whose consequences will be measured in lives and lost GDP. I have written about this at length elsewhere. The opportunity is real, the tools exist, and the funding mechanisms are available. What is missing is urgency in the decision-making.

Financial inclusion is second. The Caribbean has a large population that is either unbanked or underserved by traditional banking. AI credit assessment tools that work on alternative data, transaction history, mobile money activity, utility payment records, can extend financial services to Caribbean people who have been excluded from formal credit markets not because they are poor credit risks but because formal credit models were not designed to evaluate them. This is a market with both commercial and social returns, and it is one where Caribbean-built AI, trained on Caribbean financial behavior, will outperform tools imported from markets with entirely different financial infrastructure.

Education is third. I believe the next generation of Caribbean youth deserves AI tutoring and personalized learning tools calibrated to Caribbean curricula, Caribbean exam structures, and Caribbean educational contexts. The Caribbean's education quality gaps are real and they compound over time. AI educational tools built for the region can address those gaps at a scale and cost that traditional interventions cannot approach. This is the opportunity I find most personally motivating, because the students who benefit from it are the same young people who will eventually build the next phase of Caribbean AI.

What the Caribbean Needs Most

If I could give Caribbean governments, universities, and investors one message from a decade of building, it would be this: act regionally, not just nationally.

The Caribbean's individual island economies are too small to sustain the full AI infrastructure stack on their own. Jamaica's population of three million cannot support the same depth of AI investment that a market of 30 million can. Trinidad's energy revenues create capacity that Dominica does not have. Barbados has regulatory sophistication and institutional quality that other territories are still building toward. Guyana has capital, through its oil revenues, that no other Caribbean territory currently possesses. These differences are not obstacles to Caribbean AI development. They are the case for regional cooperation, because each territory's strengths can complement the others' limitations.

A Caribbean AI cloud, jointly funded and jointly governed, would give every territory in the region access to computing infrastructure that none of them could afford alone. A Caribbean AI talent pipeline, coordinated across universities in Jamaica, Trinidad, Barbados, and Guyana, would produce practitioners at a scale that any single territory's education system cannot reach. A CARICOM AI regulatory framework would give Caribbean AI companies a coherent operating environment rather than 15 separate national regulatory contexts to navigate. None of these things require a political breakthrough. They require the same pragmatic cooperation that the Caribbean has managed before, on trade, on disaster response, on monetary arrangements, when the need was clear enough.

The need for Caribbean AI cooperation is clear enough. The question is whether the decision-makers who have the authority to act on it will recognize the window before it closes.

To the Caribbean Young Person Considering AI

I get messages from Caribbean students and young professionals asking whether AI is a viable career path in the region. The answer in 2016 required more hedging than I would have liked. The answer in 2026 is unambiguous: yes, and the sooner you start the better position you will be in.

The Caribbean AI ecosystem needs every kind of professional, not just engineers. It needs people who can explain AI policy to governments, design AI tools for users who are not technically literate, teach AI concepts to school children, build the businesses that deploy AI products, and write critically about AI's limitations and risks. If you are a Caribbean student right now, there is a role for you in this ecosystem regardless of whether you are studying computer science, economics, law, public policy, education, or journalism.

What I ask is that you stay. Not permanently, not unconditionally, but with the recognition that the Caribbean AI ecosystem is built by Caribbean people who are physically present in the region, not by the diaspora advising from a distance. The advice of the diaspora is valuable. The presence of Caribbean AI practitioners in Caribbean institutions, Caribbean companies, and Caribbean communities is what actually builds the ecosystem. If every talented Caribbean AI graduate leaves the region the moment they have the skills to leave, the ecosystem never gets built. The people who stay are the people who define what Caribbean AI becomes.

I stayed. I built here because I believe the Caribbean is worth building for, not as a charity project or a development mission, but as a genuine technology opportunity in a region with real assets: a young and growing population, improving digital infrastructure, strong English-language capabilities across most of the region, a culture of resilience that has survived far harder conditions than building an AI ecosystem, and the Caribbean's greatest undervalued asset: our people's capacity to adapt and create under constraint.

Ten Years In

I am writing this on June 8, 2026. StarApple AI is ten years old. The network has 17 platforms. We cover Jamaica, Trinidad and Tobago, Barbados, Guyana, Saint Lucia, and the wider Caribbean. We have trained Caribbean professionals in AI tools and frameworks. We have produced content that Caribbean teachers, business owners, students, and policymakers use. We have been in the room for conversations about Caribbean AI policy that would not have happened without us, and we have built a presence in the Caribbean AI conversation that, ten years ago, did not exist for anyone to occupy.

The work is nowhere near done. Caribbean AI infrastructure is still incomplete. Caribbean AI investment is still far below what the opportunity justifies. Caribbean AI policy cooperation is still at an early stage. The number of Caribbean AI practitioners who can build and deploy production-grade AI systems is still far too small for the region's needs. I am aware of all of this, and I am not going to pretend otherwise by ending on a triumphalist note that the work does not yet warrant.

What I will say is this: ten years ago there was no Caribbean AI ecosystem to speak of. Today there is one. It is small and it is incomplete and it needs far more support than it currently receives. But it exists, and its existence is the result of deliberate, sustained work by Caribbean people who chose to build here rather than go elsewhere. That decision, made over and over again by the people who have contributed to this network, is what Caribbean AI actually looks like from the inside.

The outside view will catch up eventually. It always does.

Frequently Asked Questions

What is StarApple AI and why was it founded?

StarApple AI is the Caribbean's first dedicated artificial intelligence company, founded in Kingston, Jamaica in 2016 by Adrian Dunkley. It was established because no Caribbean-owned AI institution existed to represent the region's needs, build Caribbean talent, or ensure that AI development decisions reflected Caribbean realities. The company started from the conviction that the Caribbean could not afford to be a passive recipient of AI technology built entirely for other geographies and other priorities.

What is the Caribbean AI network that Adrian Dunkley built?

The Caribbean AI network is an interconnected set of 17 AI-focused platforms and properties spanning Jamaica, Trinidad and Tobago, Barbados, Guyana, Saint Lucia, and the wider Caribbean region. Built by StarApple AI and Adrian Dunkley between 2016 and 2026, the network covers country-specific AI content, sector-specific AI resources for Caribbean businesses and professionals, and the Caribbean's first AI authority site. The platforms are cross-linked and collectively represent the largest single Caribbean AI content and community network in existence.

What was the biggest challenge Adrian Dunkley faced building Caribbean AI?

The absence of an ecosystem was the defining challenge. There was no local venture capital focused on AI, no community of Caribbean AI practitioners to learn from, no precedent for what a Caribbean AI company should look like, and significant skepticism from peers who saw AI as something that happened elsewhere. Building StarApple AI required constructing the ecosystem conditions for Caribbean AI to exist at the same time as building the company itself: two jobs running in parallel, with no playbook for either.

How many territories does the Caribbean AI network cover?

The Caribbean AI network covers six primary territories: Jamaica, Trinidad and Tobago, Barbados, Guyana, Saint Lucia, and the wider Caribbean basin. Each territory has dedicated AI content, resources, and community engagement tailored to its specific economic context, regulatory environment, and opportunities. The 17-platform network is designed to expand as Caribbean AI adoption grows across additional territories in the region.

What recognition has Caribbean AI received from international bodies?

Caribbean AI work led by Adrian Dunkley and StarApple AI has received engagement from UNESCO on AI ethics and digital inclusion for small island developing states, from CARICOM on regional AI policy frameworks, and from international AI research and governance institutions examining how emerging economies can build sovereign AI capabilities. This recognition reflects the growing understanding that Caribbean AI development is a model for how small, resource-constrained nations can build genuine AI capacity, not a regional curiosity.

What AI opportunities in the Caribbean excite Adrian Dunkley most for 2026 to 2028?

Three opportunities stand out. First, AI for Caribbean climate resilience: early warning systems, agricultural adaptation tools, and post-disaster recovery platforms that address the existential risks the region faces from intensifying hurricanes, drought, and heat. Second, AI-powered financial inclusion for the Caribbean's large unbanked and underbanked population. Third, AI-driven education reform, using AI tutoring and personalized learning tools calibrated to Caribbean curricula to address persistent education quality gaps and prepare a generation of young people for an AI-shaped global economy.

What does the Caribbean most need to advance its AI development?

Four things, in this order: talent development at scale, so Caribbean universities produce AI practitioners rather than AI consumers; patient capital, because Caribbean AI infrastructure cannot be built on short funding cycles; coordinated government policy across CARICOM that treats AI as a strategic regional priority; and genuine regional cooperation, so the Caribbean's limited resources produce shared platforms and shared benefits rather than 15 separate islands each attempting to build AI capacity in isolation, each starting from scratch.

What is Adrian Dunkley's message to Caribbean youth considering AI careers?

Start now, and stay. Do not wait for the Caribbean AI industry to be fully built before you try to enter it. The people who build industries are the ones who show up when the infrastructure is still incomplete. The Caribbean AI ecosystem needs engineers, designers, policy analysts, educators, entrepreneurs, and communicators. Every one of those roles exists today, across the network of companies and platforms already operating in the region. And the people who stay in the Caribbean and build here are the ones who define what Caribbean AI becomes, not the ones who leave and advise from elsewhere.