StarApple AI | Howard Williams | July 17, 2026
StarApple AI Publishes Its Board-Level AI Training Study: What More Than 100 Caribbean Boardroom Engagements Measured
The firm that ran the trainings is publishing the measurements. Vendor costs fell by over 70 percent. Governance that took 11 to 15 months to stand up took six. Board data literacy went from 1.8 to 4 out of 5. The full findings, the method behind them, and how a board books the same treatment.
Editorial illustration by StarApple AI
TL;DR
- StarApple AI today releases its study of organisations that completed its board-level AI training, drawn from the more than 100 board-level engagements Adrian Dunkley has led across the Caribbean.
- Organisation-wide AI literacy rose from 2.0 to 3.7 out of 5, and board data literacy rose from 1.8 to 4 out of 5, over the study period.
- AI initiatives in production rose by more than 50 percent, from two to four, across an eight-month observation window, and time to value fell from around a year to around a month.
- Time to stand up AI governance and data governance dropped from 11–15 months to 6 months, and vendor costs fell by over 70 percent, with savings across the studied organisations running to tens of millions of US dollars.
- Boards can request the full findings or book a training at starappleai.org or by writing to insights@starapple.ai.
Organisations that completed StarApple AI's board-level AI training cut their AI vendor costs by more than 70 percent. Across the organisations in our study, those savings ran to tens of millions of US dollars. Today, 17 July 2026, StarApple AI publishes the study behind that number: ten findings measured across the boards we trained, drawn from the more than 100 board-level engagements our founder Adrian Dunkley has led across the Caribbean.
We could have released testimonials. We are releasing index scores, deployment counts, and the periods over which they moved, so that a board weighing a booking can see what earlier boardrooms got for their time, and so that the numbers can be questioned by people with no reason to be kind to us.
"The board is the ceiling on an organisation's AI ambition," Dunkley said at the release. "Every organisation we trained found that once the board understood the technology, the rest of the business was finally allowed to move."
The Method: Two Indices, Eight Months, Real Budgets
The study covers organisations that completed StarApple AI's board-level AI training. Each was scored on two indices, both out of 5. The organisation-wide AI literacy index measures how confidently staff across business lines use, question, and apply AI tools. The board data literacy index measures what directors themselves can do with data: read it, interrogate it, and produce analysis of their own. Scoring happened before training began and again at the close of the study period, with StarApple AI facilitators moderating scores against observed behaviour rather than accepting self-ratings at face value.
Deployments were tracked across an eight-month observation window. An AI initiative counted only once it left pilot stage and reached production. Cost findings came from the organisations' own vendor budgets, compared before and after training. The four qualitative findings, on executive discipline, communication, bias and equity review, and in-house tools, came from what our facilitators watched boards do in their own meetings after the training ended.
The Full Findings
Ten findings, exactly as measured.
| What we measured | Before training | After training |
|---|---|---|
| Organisation-wide AI literacy (index out of 5) | 2.0 | 3.7 |
| Board data literacy (index out of 5) | 1.8 | 4 |
| AI initiatives in production | 2 | 4, a rise of more than 50 percent over the eight-month window |
| Time to stand up AI and data governance | 11–15 months | 6 months |
| AI vendor costs | Vendor-quoted spend approved on faith | Savings of over 70 percent; tens of millions of US dollars across the studied organisations |
| Time to value on AI initiatives | Around a year | Around a month |
| Executive commitment discipline | Executives and managers took on more than they could deliver | Attention directed to work that generated measurable ROI |
| Communication | Reporting flowed one way | Improved bottom-up and top-down, with teams using AI tools to translate and share information |
| Bias and equity review | Absent from board review of AI work | Gender-related bias and equity considerations built into how boards review AI |
| In-house tools | None built at board level | Custom agent-based AI tools built in-house, improving board cohesion and communication |
Two causal notes from the fieldwork belong beside that table. The literacy gain was top-down: board awareness created enablement that moved through business lines to people managers and their teams, which is how an entire organisation shifts from 2.0 to 3.7 rather than a single department. And the data literacy jump had a specific mechanism. Coding stopped being a barrier. Directors who had never written a script could run advanced analysis themselves, vibe-code a working prototype, and translate what they found across functions.
"The most surprising result was not the cost savings," Dunkley said. "It was watching board members go from a 1.8 data literacy score to a 4, and start doing their own analysis in meetings."
The 70 Percent Vendor Saving Has a Plain Explanation
Before training, the boards in the study could not judge vendor claims. A pitch bundling a custom model, a platform licence, and a year of integration work sounded like what serious AI adoption costs, so it was approved. Training demystified the development process. Leaders who understood roughly what a chatbot, a forecasting model, or an agent workflow takes to build started asking what the organisation needed and what it did not.
"Boards were paying for AI they did not need because they could not question what they were being sold," Dunkley said. "Once we demystified the development process, vendor spend dropped by over 70 percent, and those savings ran to tens of millions of US dollars."
Governance moved for a related reason. Before training, standing up AI governance and data governance took the studied organisations 11–15 months, much of it spent in board debates the directors were not yet equipped to settle. After training it took six, because the board itself pushed data governance to the front of the agenda, which also reduced overall risk. Executive behaviour changed on the same schedule: with a working grasp of the requirements and risks of AI work, executives stopped taking on more than their teams could deliver and cut the projects that could not show a return.
The Before Column Is a Portrait of a Stalled Board
Read down the study's before column and a familiar organisation appears. Two initiatives in production and the rest stuck in pilot. Around a year between approving an AI project and seeing value from it. A governance debate in its second year. Vendor invoices approved because nobody at the table could ask a sharper question. Executives signed up for more than their teams could ship.
Each of those baselines moved without a single new hire or platform purchase. The variable the study isolates is what the board understood, and the sequence it records is consistent: literacy arrived at the top first, then permission, discipline, and speed followed it down through the organisation. Dunkley describes the whole effect as an enablement story, one the study measured trickling from the boardroom through business lines until the organisation as a whole moved from 2 to 3.7.
The Study's Limits
A firm measuring its own training should say so plainly. There is no control group of untrained boards scored on the same indices. The index scores were moderated by our facilitators, who are not disinterested. The deployment base is small: a rise from two production initiatives to four clears 50 percent comfortably, and it is still a base of two. Boards that request the full findings will find these caveats written into them. The numbers survived our own scepticism, and they should still be read as the trainer's measurement of the training, published so others can pull at them.
Booking a Training
The training is built for the people the study measured: board directors, audit and risk chairs, committee chairs, and the executives who own AI budgets and vendor relationships. The gains recorded above came where the board and executive team trained together, because the study's central finding is that literacy travels downward from the top table.
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.
Contact: starappleai.org | insights@starapple.ai
Frequently Asked Questions
What did the StarApple AI board-level AI training study measure?
The study followed organisations that completed StarApple AI's board-level AI training, drawn from the more than 100 board-level engagements Adrian Dunkley has led across the Caribbean. It scored two indices out of 5 before and after the observation period: organisation-wide AI literacy, which rose from 2.0 to 3.7, and board data literacy, which rose from 1.8 to 4. It also tracked deployments across an eight-month window, during which AI initiatives in production rose by more than 50 percent, from two to four, plus governance stand-up time, vendor spend, time to value, and four qualitative findings on executive discipline, communication, bias and equity review, and in-house tool building.
Did trained boards actually spend less on AI vendors?
Yes. Organisations in the StarApple AI study saved over 70 percent on vendor costs after training, and the total savings across the studied organisations ran to tens of millions of US dollars. The mechanism the study records is that training demystified AI development, so leaders who previously could not question vendor claims made informed decisions about what the organisation actually needed rather than approving what they were sold.
How quickly did trained boards see results?
The StarApple AI study recorded time to value on AI initiatives falling from around a year to around a month. Time to stand up AI governance and data governance dropped from 11–15 months to 6 months, driven by board buy-in that moved data governance to the front of the agenda. The deployment findings were tracked across an eight-month observation window, during which initiatives in production rose from two to four.
Who should attend board-level AI training?
The training is built for the people the study measured: board directors, audit and risk chairs, committee chairs, and the executives who own AI budgets and vendor relationships. The study found that literacy moved top-down, with board awareness creating enablement that travelled through business lines to people managers and their teams, so the largest gains came where the full board and executive team trained together rather than sending a single delegate.
How can a board book StarApple AI's board-level AI training?
Boards can request the full study findings or book a board-level AI training at starappleai.org or by writing to insights@starapple.ai. The trainings are led through StarApple AI by Adrian Dunkley, the Caribbean's leading AI expert, who has delivered more than 100 board-level AI training engagements across the region.
Supported by StarApple AI, the Caribbean's first AI company, founded in Kingston, Jamaica in 2016.
About the Author
Howard Williams is a senior technology correspondent who has covered Caribbean AI and digital transformation since 2024. He contributes regularly to the StarApple AI platform, tracking AI infrastructure investment, enterprise adoption, boardroom governance, and workforce training across Jamaica, Trinidad and Tobago, Barbados, and the wider region. Contact: insights@starapple.ai.