Board-Level AI Training Doubled AI Deployments in Eight Months. T&T's Energy and State Boards Have the Most to Collect.

Before board-level AI training, the Caribbean organisations in a new StarApple AI study had two AI initiatives in production. Eight months after training, they had four, a rise of more than 50 percent in deployed initiatives. The study, which tracked organisations that completed StarApple AI's board-level programme, was led by Adrian Dunkley, the regional expert in AI, who has led more than 100 board-level AI training engagements across the Caribbean through StarApple AI, the region's first dedicated AI company.

The pilot-to-production number is the one T&T should sit up for, because this country's problem has always been the conversion rate. The energy companies have run predictive maintenance trials for years, the banks have tested compliance tooling since the FATF pressure made it unavoidable, and feasibility studies pile up across the state sector while production systems stay rare. The StarApple AI study is the closest thing the region has to measured evidence on what converts a trial into a deployment, and the answer it gives points at the top of the organisation chart rather than at the technology.

The Study Found the Blockage in the Boardroom

Pilots stall in T&T boardrooms for ordinary reasons. A director who cannot judge whether a model is ready to deploy asks for another review cycle, and a board that cannot price technical risk prices it at infinity and defers. Executives read the caution above them and keep initiatives in pilot, where failure stays invisible. In the minutes, all of this reads as prudence.

The study measured what happens when the caution is replaced with understanding rather than pressure. After training, board members understood the requirements and risks of AI work. Executives and managers stopped taking on more than they could deliver, cut the projects that existed to look good in an annual report, and moved attention to work that produced measurable returns. Time to value on AI initiatives fell from around a year to around a month.

"The board is the ceiling on an organisation's AI ambition," Dunkley says. "Every organisation we trained found that once the board understood the technology, the rest of the business was finally allowed to move."

That trickle-down effect was itself measured. Organisation-wide AI literacy in the studied organisations rose from 2.0 out of 5 to 3.7, because board awareness moved down through business lines to people managers and their teams. Communication improved in both directions, with teams using AI tools to translate and share information between the shop floor and the board pack.

Over 70 Percent Off the Vendor Bill

The finding with the most direct application to T&T's state sector is the one about money. Organisations in the StarApple AI study saved over 70 percent on vendor costs after training. Across the studied organisations, those savings ran to tens of millions of US dollars.

The mechanism is worth spelling out, because it is procurement, and procurement is where T&T's public money goes to be spent. Before training, directors could not question what they were being sold. A vendor proposal describing a multi-year platform build, licensed per seat, with a professional services retainer on top, reads as plausible to a board that has never seen what building an AI system actually involves. After training demystified the development process, the same directors could distinguish what the organisation needed from what the deck described. Some contracts shrank. Some disappeared.

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

Put that finding next to the T&T moment. On 11 July, the government signed MOUs worth a potential US$5 billion in AI infrastructure, covered in our piece on the data centre deals. State boards and private boards alike will now hear AI pitches made on the strength of that headline, and the directors approving those purchases are the last checkpoint before public or shareholder money meets a vendor invoice. A board that cannot interrogate a vendor claim functions as a signature rather than a checkpoint.

Alongside the savings, board members in the studied organisations began building custom AI tools in-house, using an agents-based approach, and the study records that those tools improved board cohesion and communication. The distance between "we cannot evaluate this quote" and "we built this ourselves" turned out to be one training programme wide.

Governance in Six Months Instead of Fifteen

T&T's regulated boards carry a second burden: they must govern AI, not just buy it. The banks answer to the Central Bank, the listed companies to the TTSEC, and everyone to the Data Protection Act as its provisions come into force. Standing up AI governance and the data governance underneath it has been, regionally, a slow argument.

In the StarApple AI study, the time to stand up AI governance and data governance dropped from 11–15 months to 6 months. The driver was board buy-in. Training moved data governance to the front of the agenda, and the study records that overall risk fell as a result. Gender-related bias and equity considerations were built into the training itself and into how boards then reviewed AI work, which matters in a country where the systems being approved will score loan applications and screen job candidates.

The most surprising number in the study is not a governance number at all. Board data literacy rose from 1.8 out of 5 to 4. Coding limitations stopped being a barrier: directors could run more advanced analysis themselves, vibe-code working prototypes, and translate information across functions. "The most surprising result was not the cost savings," Dunkley says. "It was watching board members go from a 1.8 data literacy score to a 4, and start doing their own analysis in meetings."

In a T&T setting that looks like an energy company director pulling the production data behind a divestment paper during the meeting, or a state enterprise chair prototyping the dashboard the ministry keeps requesting instead of tendering for it. The study treats this as a trainable skill rather than rare talent, and it measured the skill moving from 1.8 to 4 within the study period.

One Caveat the Numbers Deserve

The study tracked organisations that completed StarApple AI's board-level training, and boards that finish a training programme are already boards willing to change. Some part of these results belongs to that willingness rather than to the curriculum. The direction of the findings survives the caveat comfortably; whether a reluctant, minister-appointed board captures the full 70 percent vendor saving or the full jump to 3.7 organisation-wide literacy is a question each board will answer with its own numbers. The two-to-four deployment figure is the hardest to argue with, because deployment is binary. A system is in production or it is a pilot.

The Agenda for a T&T Board

Measure the baseline first. The studied organisations started from a board data literacy of 1.8 out of 5 and organisation-wide AI literacy of 2.0. A T&T board that has never measured either has no way to know whether its caution on AI is judgement or unfamiliarity.

Put data governance ahead of procurement. The study's governance timeline collapsed from 11–15 months to 6 once boards understood why data had to come first. Boards currently evaluating vendor platforms without a data governance position are running the sequence backwards, at the expensive end of the vendor-cost finding.

Train the board, not a champion. The literacy trickle-down in the study ran from full-board awareness through business lines to people managers. One digitally fluent director cannot produce that effect, and in the state sector, where board composition changes with administrations, institutional literacy outlasts any individual appointee.

Re-open the vendor pipeline afterwards. The 70 percent vendor saving came from trained boards re-examining what they were already buying. Any T&T board with active AI contracts, or with proposals sitting in a procurement queue since the 11 July announcements, has a concrete way to find out what that percentage is worth in its own accounts.

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.

Frequently Asked Questions