The Boardroom Is the Bottleneck: A StarApple AI Study Makes the Caribbean Case for Board-Level AI Training

Caribbean organisations that put their boards through StarApple AI's board-level AI training stood up AI governance and data governance in 6 months. Before training, the same organisations were taking 11–15 months. That finding comes from a StarApple AI study of organisations that completed the programme, led by Adrian Dunkley, who has run more than 100 board-level AI training engagements across the region. Since the tools on the market stayed the same across the study period, the acceleration comes down to one variable: what the directors in the room understood.

Regional AI debate spends most of its energy on compute, connectivity, talent pipelines and which tools to buy. The study points somewhere else. Every constraint it measured, from stalled pilots to inflated vendor bills to governance arguments that ran past a year, loosened once the board was educated, and none of them required new infrastructure to fix. For CAIA, which has spent two years arguing that the region's AI gap is institutional rather than technical, this is the first Caribbean dataset that puts numbers behind the claim.

The Study in Numbers

• Organisation-wide AI literacy rose from 2.0 out of 5 to 3.7 out of 5 over the study period, as board awareness moved down through business lines to people managers and their teams.
• Board data literacy rose from 1.8 out of 5 to 4 out of 5, with directors running their own analysis and building working prototypes.
• Time to stand up AI governance and data governance fell from 11–15 months to 6 months, driven by board buy-in.
• Deployed AI initiatives rose by more than 50 percent, from two to four, over eight months.
• Vendor costs fell by over 70 percent, with total savings across the studied organisations in the tens of millions of US dollars.
• Time to value fell from around a year to around a month.

Literacy Moved From the Boardroom Down

The result with the widest regional consequence is the one that happened furthest from the boardroom. Across the studied organisations, the internal AI literacy index rose from 2.0 out of 5 to 3.7 out of 5, and the study traces the mechanism directly: board awareness gave the rest of the organisation permission to move, and that permission travelled down through business lines, to people managers, and on to their teams. Staff became literate because the people who set budgets and sign policies stopped treating AI as something to be feared or delegated, and the room to learn flowed downhill with the budgets.

"AI literacy at the top is an enablement story. We measured it trickling down from the boardroom through business lines to people managers, and the whole organisation moved from a 2 to a 3.7," Dunkley said.

The boards themselves moved further still. Board data literacy rose from 1.8 out of 5 to 4 out of 5 over the study period. Coding stopped being a barrier: directors ran more advanced analysis themselves, vibe-coded working prototypes, and translated information across functions instead of waiting for a technical team to interpret it for them. Several boards went on to build custom AI tools in-house using an agents-based approach, which the study credits with improving board cohesion and communication. Communication improved in both directions across the wider organisation too, bottom-up as well as top-down, with teams using AI tools to translate and share information between levels that had previously talked past each other.

For a region of small states, the trickle-down finding matters more than any single organisational result. CARICOM governments cannot train every worker directly, and national literacy campaigns are slow and expensive. Training the few hundred people who sit on the region's most consequential boards is neither.

Governance in Six Months Instead of Fifteen

The governance finding is the one regulators should sit with. Standing up AI governance and data governance had been taking the studied organisations 11–15 months, and the study is clear about where that time went: it was consumed by the internal argument over whether governance mattered at all, an argument that only exists when the people at the top cannot evaluate what they are being asked to approve. After training, the same task took 6 months. Training moved data governance to the front of the agenda and reduced overall risk, because boards finally understood why the data layer had to be governed before the model layer could be trusted.

"Governance went from a 15-month argument to a 6-month build. Nothing about the technology changed. What changed was that the board understood why data governance had to come first," Dunkley said.

Caribbean supervisors already treat director competence as a regulatory matter in banking and insurance, where fit-and-proper assessments ask whether a board can actually oversee the risks on its balance sheet. The study gives regulators, central banks and data protection commissioners a measured case for extending the same logic to AI: a board that cannot question a model cannot govern one, and the cost of that gap is now quantified in months of delay and millions in spend. It also gives the region's directors' institutes a concrete curriculum argument at exactly the moment governments from Kingston to Port of Spain are writing national AI policies that will need governed institutions to land in.

Equity Written Into the Board's Review

One finding takes a single sentence in the study and deserves more attention than it will get. Gender-related bias and equity considerations were built into the training itself, and the study found they were then built into how boards reviewed AI work afterwards. Directors who had been taught to ask who a model fails for kept asking after the training ended, in procurement decisions and project reviews. For a region where AI systems will increasingly mediate credit, hiring and public services, board-level equity scrutiny is the cheapest safeguard available, and this study shows it can be installed deliberately.

Discipline, Vendor Spend and the Speed of Value

The remaining findings describe what an educated board does with its new judgement. Executives and managers stopped taking on more than they could deliver. Vanity projects were cut and attention went to work that generated measurable ROI, which shows up in the deployment numbers: initiatives that left pilot stage and reached production rose by more than 50 percent, from two deployed initiatives to four, over eight months. Time to value fell from around a year to around a month.

Then there is the money. Organisations in the study saved over 70 percent on vendor costs after training, because leaders who had previously been unable to judge vendor claims could suddenly tell the difference between what they were being sold and what they needed. Across the studied organisations those savings ran to tens of millions of US dollars, capital that stayed in Caribbean balance sheets instead of leaving the region as licence fees for tools nobody had scoped.

"Boards were paying for AI they did not need because they could not question what they were being sold. Once we demystified the development process, vendor spend dropped by over 70 percent, and those savings ran to tens of millions of US dollars," Dunkley said.

The study has a limit worth stating. It covers organisations that chose to buy board training, and organisations willing to invest in their directors may be the ones already inclined to move well. The study cannot fully separate the effect of the training from the ambition of the buyer. What it can show is that every measured constraint moved in the same direction at the same time, and that the trigger in each case was the board.

Next Steps for Caribbean Boards and Regulators

CAIA draws four working conclusions from the study. Boards should treat AI education as a governance obligation on par with financial literacy, scheduled and assessed rather than optional. Directors' institutes and professional bodies across CARICOM should put board-level AI training into their certification pathways now, while the evidence is fresh. Regulators should start asking regulated entities how their boards are educated on AI, in the same breath as they ask about cyber risk. And governments drafting national AI strategies should budget for director education alongside workforce programmes, because the study suggests board education is what makes workforce programmes land.

Adrian Dunkley, the Caribbean's leading AI expert, has led more than 100 board-level AI training engagements through StarApple AI. Boards can request the full study findings or book a training at starappleai.org or by writing to insights@starapple.ai.

Disclosure: Adrian Dunkley founded the Caribbean AI Association in 2024 and leads StarApple AI, which conducted the study reported here. CAIA states that connection plainly and reports the figures as the study presents them.