Today, 12 September 2026, thirty-five women in Barbados sit down for the first of three Saturday sessions in a programme called Build a Business With AI. The National Organisation of Women (NOW) and Amplifi AI Inc, a Barbados-based AI firm, launched it on 1 September. Participants pay nothing, although the organisers value the training, mentorship and platform access at more than $20,000. Sessions run on 12, 19 and 26 September, and the cohort pitches to a panel and audience on 10 October.
The cohort is small. The problem it points at is large: the Inter-American Development Bank puts the financing gap for women-owned micro, small and medium enterprises across Latin America and the Caribbean at $93 billion. AI tools can help with the preparation side of that gap, meaning market research, pricing and financial projections. They do not change collateral rules or the interest rates women-led firms are offered, and nothing in the evidence so far suggests otherwise.
What the three Saturday sessions cover
The curriculum starts with the business and brings in the software afterwards. Participants work through customer identification, market validation, pricing and revenue potential, then apply AI tools at each stage: testing a value proposition against a defined customer segment, or modelling what a price change does to margin.
Amplifi AI chief executive Janelle Germain explained the order this way: "This technology makes foundational knowledge more important, not less." I agree with her from experience. A founder who cannot explain her own unit economics will get an AI-generated business plan that looks polished and contains numbers nobody can defend.
NOW president Melissa Savoury-Gittens said the programme grew out of an earlier NOW course in digital marketing for young people, after people outside that age group asked for something similar. Her summary of the aim: "Economic empowerment means access. It means skills. It means knowledge."
Applications ran through Amplifi AI's website and closed about a week after launch. Thirty-five places in a country of roughly 267,000 people is a small sample. Three weeks from announcement to first session, and six weeks to a pitch day, is a faster cycle than most public skills programmes in the region manage.
The financing gap in numbers
The data behind the programme comes from three sources, and the figures are worth keeping together:
| Measure | Figure | Source |
|---|---|---|
| Financing gap, women-owned MSMEs, Latin America and the Caribbean | $93 billion | IDB |
| Adults over 15 who have borrowed from a financial institution | 24% of women, 35% of men | Regional survey data, Latin America and the Caribbean |
| Caribbean women entrepreneurs who are unserved or underserved | Up to 73% | ICR Facility access-to-finance study |
| Women's share of formal MSME ownership in Barbados | 40.8%, up from 34.4% in 2016 | Barbados Small Business Association |
| Women's share of Barbadian businesses started 2020–2025 | About 58% | Barbados Small Business Association |
| Jamaica's absolute gap for women-owned firms | $2.717 billion, the largest in dollar terms among territories studied | ICR Facility |
| Belize's gap as a share of GDP | 26%, the highest of territories studied | ICR Facility |
Barbadian women are starting businesses faster than men. Growing them is where the trouble starts. The Small Business Association's figures show the constraint arriving after start-up, when a firm needs commercial space, staff, equipment and compliance work, all of which need capital.
Where AI tools save a solo founder money
A founder with no accountant, no marketing staff and no lawyer on retainer gets the most from these tools, in three places.
- Market validation. A research firm charges thousands of dollars for a customer segmentation study. An AI research assistant, prompted by someone who knows her customers, produces a rough version in an afternoon. It will be wrong in places. It is right often enough to sharpen a pitch, provided every figure is checked against a real source.
- Pricing. A founder can model three price points against estimated volume and margin in the time one spreadsheet used to take. A workable prompt: "I sell handmade soap in Bridgetown at BBD 12 a bar, cost BBD 5.50, about 300 bars a month. Show margin and monthly profit at BBD 10, 12 and 15, assuming volume falls 10% for every BBD 2 increase. Show the arithmetic."
- Financial projections. Inconsistent cash-flow projections are one of the most common and most fixable reasons a loan application stalls. AI tools will not invent revenue, but they catch arithmetic errors and growth assumptions that no loan officer will believe.
- Pitch rehearsal. Ask the tool to play a sceptical loan officer and question the plan line by line. It costs nothing and it finds the weak answers before the panel does.
Institutional work is happening at the same time. The University of the West Indies and Sagicor opened an AI and Financial Services Hub in Barbados in May 2026, backed by a $5 million regional investment in applied research across the financial sector. The hub works with banks and insurers. NOW's cohort works with individual founders. Both assume the region can use tools that already exist elsewhere, applied to Caribbean numbers and Caribbean lenders.
The Core Argument
AI training gives a founder with no staff the market research, pricing models and financial projections a funded startup would pay a team to produce. That narrows the preparation gap. The collateral gap stays where it is, and a programme that claimed otherwise would lose the trust these cohorts earn.
What training cannot change: collateral, credit history and risk pricing
IDB Invest and the ICR Facility have both documented that women-led businesses in the Caribbean are more likely to be denied credit than comparable male-led firms. When approved, they tend to get smaller loans, shorter repayment periods, higher interest rates and heavier collateral requirements. A better spreadsheet changes none of those terms.
A stronger application does help at the margin. A loan officer looking at a defensible market analysis has one less reason to decline. The barriers the IDB and ICR Facility describe sit elsewhere: collateral a founder does not own, a credit history a first-generation entrepreneur has not had the chance to build, and risk pricing that treats women-led firms as a category instead of judging each firm on its numbers.
This is also where my own argument is weakest. I am assuming better preparation shows up in approval rates, and there is no Caribbean data yet that tests it. Nobody has tracked a cohort like this one through to loan decisions.
How other territories could copy the model
The format needs three things: a national women's organisation that can recruit, a local AI firm willing to donate teaching time, and a sponsor for pitch-day prizes and venue costs. None of that requires a $5 million hub. A government agency or a development bank office in Kingston, Port of Spain or Georgetown could fund a version on a small budget.
What should travel with the format is NOW's candour about scope. Pitch the programme as preparation, measure it on financing outcomes, and leave collateral, credit history and risk pricing to the development finance institutions, regional banks and financial regulators who control them.
What to do next
- Founders in Barbados: before the next cohort opens, write down your unit economics (price, cost per unit, monthly volume) on one page. Then run the pricing prompt above with your own numbers and check every result by hand. It takes about an hour and costs nothing on a free AI tier.
- Founders applying for credit: ask your bank or credit union for its checklist before you draft anything, and use an AI tool to test your plan against that list, section by section. Treat any figure the tool produces without a named source as a placeholder until you have verified it.
- NOW and Amplifi AI: publish a follow-up six and twelve months after the 10 October pitch day, covering how many of the 35 applied for financing, how many were approved, and on what terms. I know of no Caribbean training programme that has published outcome data of this kind.
- Women's organisations in Jamaica, Trinidad and Tobago, Guyana and Belize: approach one local AI firm and one sponsor with a three-session, six-week plan modelled on this one, capped at about 35 places so mentors can give each founder real attention.
- Lenders and development finance institutions: offer cohort graduates a named contact and a written reason for every decline. Declines with reasons are the data this region needs to separate plan quality from collateral and pricing barriers.
Frequently Asked Questions
Can women outside Barbados apply for Build a Business With AI?
Not for the September 2026 cohort, which had 35 places and was recruited through NOW in Barbados. Applications went through Amplifi AI Inc's own website, so that is the place to watch for future intakes. Neither organisation has said publicly whether later cohorts will be open to women in other territories, so ask them directly before planning around it.
How do I check market figures an AI tool gives me before they go into a business plan?
Ask the tool to name the source for every figure, then open that source yourself. AI assistants regularly produce plausible statistics that do not exist, and a loan officer who finds one invented number will doubt the rest. For Barbadian population, household and trade data, the Barbados Statistical Service is the official source; for your own market, a count of real competitors and ten conversations with likely customers beat any generated estimate.
Is it safe to put customer or financial data into a free AI tool?
Keep personal data about customers out of free consumer tools unless you have checked the provider's data settings. Several providers use conversations on free tiers to improve their models unless you switch that off, and Barbados's Data Protection Act, 2019 places duties on any business that processes personal data. Use made-up or aggregated numbers when you are modelling prices and margins; the tool does not need real names or account details to do the arithmetic.
What will AI tools cost a founder after a free programme ends?
Nothing, if the free tiers are enough, and for a business plan and pricing model they usually are. At the time of writing the paid individual plans of ChatGPT, Claude and Gemini are priced at around US$20 a month, which buys longer documents, file uploads and higher usage limits. Check current pricing before committing, and pay for one tool, not three.
What should I bring to a lender along with an AI-assisted business plan?
Bring the evidence behind the plan: recent bank statements, any sales records, quotes from suppliers, and a one-page list of the assumptions in your projections with where each came from. Lenders in the region decide mainly on repayment capacity and security, so a clear statement of what you can offer as collateral, or why you cannot, saves a wasted meeting. Ask the lender for its application checklist first and match your documents to it.