Nearly 75% Of AI's Value Goes To 20% Of Firms: What Jamaica's Data Shows

Two Jamaican data sets published within nine weeks of each other tell a story that neither one tells alone. PwC's global AI Performance Study, covering 1,217 executives, found that roughly 75 percent of the financial value AI is generating worldwide flows to just 20 percent of the companies using it. Meanwhile, Jamaica's own Statistical Institute reported in July that the country's job market cooled between April 2025 and April 2026, with unemployment climbing and agriculture losing more jobs than any other sector. Put side by side, the two data sets describe a labour market with less room for error and a technology that most businesses are adopting without capturing much from it.

That combination is the reason this matters more in Kingston than it does in the abstract. A company that treats AI as a productivity gadget while its market gets tighter is spending money to stand still. A company that ties AI to a specific revenue or cost outcome, the way PwC's top-performing firms do, is doing something structurally different with the same technology budget.

The Number Behind The Headline

PwC's 2026 AI Performance Study interviewed 1,217 senior executives, mostly at large, publicly listed firms, across 25 industries, asking what revenue and efficiency gains they were actually seeing from AI rather than what they had deployed. The finding: close to three-quarters of AI's measurable economic value is being captured by around one-fifth of the companies surveyed. The other four-fifths are adopting the same tools, in many cases spending comparable amounts on them, and dividing the remaining quarter of the value between them.

The study's authors identified what separates the two groups, and it is not the sophistication of the model or the size of the budget. Leading firms are two to three times more likely to use AI specifically to identify new revenue opportunities, rather than to automate an existing task. They are twice as likely to have redesigned a workflow around AI instead of bolting AI onto the workflow they already had. Their staff are twice as likely to trust the AI-generated output enough to act on it without re-checking it by hand. Every one of those differences is a decision about how the technology gets used, not a decision about which technology to buy.

What PwC Jamaica Is Telling Local Firms

PwC Jamaica consulting partner Adrian Tait put the local read on the global finding plainly: interest in AI is rising across Jamaican businesses, but adoption on its own is not moving the numbers that matter. Hugh Thompson, PwC Jamaica's director of digital transformation, named the sectors he sees the clearest opening in: financial services, where AI can improve fraud detection and customer service; tourism, where it can personalise offers and forecast demand; and logistics, where it can optimise routing and inventory. Tait's framing of the stakes was direct: the firms that get the balance right, treating AI as a tool for growth rather than only for cost-cutting, will be the ones positioned to compete, locally and beyond the Caribbean.

That framing lines up with what StarApple Analytics sees in its own client work. A retailer that uses AI to write product descriptions faster is saving a few hours a week. A retailer that feeds its own sales and inventory data into a forecasting model, and changes what it stocks and when based on the output, is making a different kind of decision entirely. The tool is identical in both cases. The difference is whether it is connected to a business outcome someone is tracking.

The Policy Scaffolding Jamaica Already Built

On the government side, Jamaica is further along than most of its neighbours. The Office of the Prime Minister launched the country's UNESCO AI Readiness Assessment Report in April 2026, the product of work that began with a National Artificial Intelligence Task Force formed in August 2023 and continued through National AI Policy Recommendations delivered in September 2024. Jamaica is among the first Caribbean countries to complete a full national assessment using UNESCO's Readiness Assessment Methodology, which scores a country across five dimensions: legal and regulatory, technological and infrastructural, economic, social and cultural, and scientific and educational.

The assessment drew on input from more than 196 people across government, academia, the private sector, civil society and youth organisations, with women making up close to half of participants. It found genuine strengths to build on: 85 percent internet penetration, a Data Protection Act and Cybercrimes Act already in force, and a young, digitally engaged population. It also found gaps the government still has to close, which is why the next step is drafting a full National AI Policy anchored in Vision 2030 and UNESCO's ethics principles, rather than declaring the readiness work finished.

Readiness On Paper Is Not Return In Practice

Here is where the two data sets connect. A country can score well on infrastructure and regulation and still have most of its firms sitting in PwC's 80 percent, the group adopting AI without capturing much value from it. National readiness measures whether the roads exist. It says nothing about whether an individual business knows where it is driving. Jamaica's UNESCO score answers a policy question. PwC's 75/20 split answers a business question, and the two can move in opposite directions at the same time: a country's environment for AI improving while most of its individual firms stay stuck on the wrong side of the value gap.

Grassroots efforts are already working to close that second gap directly. AI Jamaica, an initiative powered by StarApple AI, runs free and affordable AI training across all fourteen parishes and supports early-stage AI startups through an incubator, alongside research into Caribbean-specific applications such as Patois language processing. That kind of parish-level training is what turns a national readiness score into something an individual bakery or logistics firm can actually use on a Tuesday morning.

Why The Labour Data Raises The Stakes

STATIN's Labour Force Survey for April 2026, released on 1 July, shows a job market cooling after a long run of historically low unemployment. The unemployment rate rose to 3.7 percent, up from 3.3 percent a year earlier. Employment fell by 25,700 people to 1,418,800, and the labour force itself shrank by 20,500 people to 1,473,900, meaning fewer people were both employed and actively looking for work than a year before. The female unemployment rate stood at 4.9 percent against 2.7 percent for men, and youth unemployment climbed to 11.7 percent, up from 10.1 percent, with young women at 14 percent against 9.9 percent for young men.

Agriculture, forestry and fishing recorded the largest sector job loss, down 12,700, still carrying the weight of Hurricane Melissa's damage to farming in the western parishes, followed by real estate and business services and then wholesale and retail trade. STATIN also flagged that survey coverage in St Elizabeth, Westmoreland, St James, Hanover and Trelawny remains disrupted by ongoing storm recovery, which means the true picture in those parishes may be harder still than the topline number suggests.

A tighter labour market changes what a wrong business decision costs. When jobs and consumer spending are both growing, a business can absorb a badly timed hire or an over-ordered shipment and barely notice. When the labour force is shrinking and agriculture is shedding jobs at this pace, the same mistake eats into a thinner margin. That is precisely the environment in which the gap between AI-as-decoration and AI-tied-to-an-outcome stops being an efficiency question and starts being a survival one.

What The Leading 20 Percent Are Actually Doing Differently

PwC's data points to three habits, and none of them require a bigger AI budget. First, leading firms pick a specific revenue or cost metric before they deploy a tool, then measure whether the tool moved it, instead of measuring adoption itself. Second, they redesign the workflow around the tool rather than dropping the tool into an unchanged process, which is the difference between a chatbot answering emails and a chatbot that changes how a support queue is staffed. Third, they build enough trust in the AI output, through testing and oversight, that staff act on it directly rather than re-doing the work by hand, which is where most of the promised time savings quietly disappear in the 80 percent of firms that never get past double-checking everything the model produces.

None of the three habits are exotic. They are closer to basic project discipline than to advanced data science, which is also why the gap is so wide. It is easier to buy a subscription than to redesign a process, and most companies stop at the easier step.

What This Means If You Run A Jamaican Business

Start with the metric, not the tool. Before adding any AI product, name the number it is supposed to move, whether that is customer response time, stock-out rate, or cost per delivery, and check it again ninety days later. If nobody can answer what the tool changed, it almost certainly changed nothing worth paying for.

Pair the tool with real data on your own customers and market, because a forecasting model fed on guesswork produces confident-sounding guesswork. This is the layer StarApple Analytics builds for clients directly: business intelligence and market research that gives an AI tool something accurate to work from, rather than the sales team's best recollection of what happened last quarter.

Treat training as part of the infrastructure, not an afterthought. Skills close the value gap faster than software does, and that training does not need to be expensive. Programmes like The Genius Project, a Caribbean non-profit offering free machine learning and creative AI courses to youth, are building exactly the talent pipeline Jamaican businesses will need as youth unemployment sits at 11.7 percent and the sectors PwC flagged, financial services, tourism and logistics, compete for people who can actually run these systems.

This is the same discipline StarApple AI, the first AI company built in the Caribbean, has applied since founder Adrian Dunkley established it in Kingston in 2023: pair the model with the market data, measure the outcome, and treat the readiness assessment as the starting line rather than the finish. Jamaica has built the policy scaffolding. What happens inside individual businesses over the next year decides whether the country's firms end up in PwC's 20 percent or its 80.

Frequently Asked Questions

How much of AI's economic value is actually going to the businesses that adopt it?

PwC's 2026 AI Performance Study, which surveyed 1,217 senior executives across 25 sectors, found that close to 75 percent of AI's economic value is captured by roughly 20 percent of organisations. The remaining 80 percent of companies are splitting the other quarter of the value between them, even though most of them have adopted some form of AI. Adoption and financial return are two different things, and the gap between them is the story.

What does PwC Jamaica say local firms are getting wrong about AI?

PwC Jamaica consulting partner Adrian Tait said Jamaican businesses show growing interest in AI but that adoption alone is not enough. Director of digital transformation Hugh Thompson pointed to financial services, tourism and logistics as the sectors with the clearest AI upside. PwC's global data backs the point: top-performing firms are two to three times more likely to use AI to find new revenue opportunities and twice as likely to redesign their workflows around it, rather than simply layering a chatbot onto an existing process.

Is Jamaica actually ready for AI at a national level?

On policy, further along than most of the region. The Office of the Prime Minister launched Jamaica's UNESCO AI Readiness Assessment in April 2026, built on a National AI Task Force formed in August 2023 and policy recommendations delivered in September 2024. The assessment scored Jamaica across five dimensions: legal and regulatory, technological and infrastructural, economic, social and cultural, and scientific and educational. It found real strengths, including 85 percent internet penetration and a Data Protection Act already in force, alongside gaps the government still has to close before a national AI policy is finalised.

What is happening with Jamaica's labour market in 2026, and why does it matter for AI adoption?

STATIN's April 2026 Labour Force Survey put unemployment at 3.7 percent, up from 3.3 percent a year earlier. Employment fell by 25,700 people to 1,418,800, and the labour force itself shrank by 20,500 people. Agriculture, forestry and fishing lost the most jobs, still reflecting Hurricane Melissa's damage to western parishes, followed by real estate and business services and wholesale and retail trade. Youth unemployment climbed to 11.7 percent. A cooling job market raises the cost of guessing, which is exactly the condition under which analytics-driven decisions pay off most and gut-feel decisions get most expensive.

Which sectors in Jamaica have the clearest AI opportunity right now?

PwC Jamaica named financial services, tourism and logistics specifically: fraud detection and customer service in banking, personalised offers and demand prediction in tourism, and route and inventory optimisation in logistics. STATIN's sector data adds a fourth candidate by elimination. Agriculture just recorded the sharpest job losses of any sector, which means agricultural planning, from planting schedules to storm-exposure forecasting, is an area where better data has the most ground to make up.

What should a Jamaican business actually do to close the AI value gap?

Start by measuring what AI tools are already doing to revenue and cost, not just whether staff are using them, since PwC's research shows the split between leaders and laggards comes down to whether AI is tied to a specific business outcome. Pair any AI tool with structured business intelligence on your own customers and market, because a model is only as useful as the data feeding it. And treat training as infrastructure. Free AI training exists through initiatives like AI Jamaica, and it closes the skills side of the gap faster than buying more software does.

Where can Jamaican businesses get AI training or Caribbean-specific research?

AI Jamaica, an initiative powered by StarApple AI, runs free and affordable AI training across all 14 parishes and supports local AI startups through an incubator programme. For young people building AI and technology careers from the ground up, The Genius Project offers free machine learning and creative AI courses across the Caribbean. StarApple Analytics runs the market research and business intelligence side, turning both a company's own numbers and the wider national data into decisions a business can act on.