- PRIIA and PRIBA confirmed on 8 September 2026 that the sixth Puerto Rico Financial Services Forum runs 2-3 November at the Sheraton Puerto Rico Convention Center in San Juan, with AI in insurance and financial crime compliance named among the agenda tracks.
- Latin America's AI-in-insurance market is valued at US$1.8 billion in 2026 and projected to reach US$8.42 billion by 2035, an 18.7% annual growth rate, per MarkWide Research, though its six-country model barely mentions the Caribbean.
- The World Bank's 2025 Global Findex counted 1.3 billion adults worldwide without a bank account; the IFC's May 2026 report on alternative data argues many can still be scored on mobile money and digital payment history, and found women borrowers perform as well as or better than men under those models.
- Being unbanked and being credit invisible are different problems. Maestro AI Labs built Credit Garden against the second one: 1.7 billion people worldwide with no usable credit history, a US$380 billion annual lending gap.
- A conference agenda cannot manufacture the structured, Caribbean-specific data a compliance or underwriting model needs to run on, the argument Maestro AI Labs has made since StarApple AI's founding in 2023.
PRIIA president Hugo Cordova called the Forum "a high-level networking destination where decision-makers connect and explore opportunities for collaboration." The sixth edition, running 2 to 3 November at the Sheraton Puerto Rico Convention Center, gives artificial intelligence in insurance and financial crime compliance its own confirmed track, sitting alongside captive insurance, correspondent banking, AML and sanctions on the published agenda. The track puts the question of what that AI runs on in front of regulators and underwriters for two days in San Juan.
What Made the November Agenda
Puerto Rico's Commissioner of Financial Institutions, Monica Rodriguez, and its Commissioner of Insurance, Suzette del Valle, both confirmed for the sixth edition, sit alongside executives from AIG, Morgan White Group, Accelerant and Banco Popular on a speaker list still filling out ahead of November. Investigative journalist Ignacio Gomez delivers the keynote. PRIBA president Eduardo Colon framed the event's purpose plainly: "The Forum continues to strengthen Puerto Rico's position as a hub for international banks and insurers." That hub role is why the agenda matters beyond Puerto Rico's own borders. PRIIA and PRIBA describe their audience as senior executives, regulators, institutional investors, insurers, international banks, compliance leaders and government officials from Puerto Rico, the United States, Latin America and the Caribbean.
The other seven confirmed tracks read like a checklist of what Caribbean financial institutions already juggle: fintech and Banking-as-a-Service, correspondent banking, captive insurance, AML and sanctions, digital assets, regulatory and law enforcement priorities, and investment strategies. AI in insurance and financial crime compliance sits alongside them as a working topic: enough of the Forum's own audience is already running AI-adjacent fraud detection, underwriting and onboarding tools that the conversation needed a room of its own.
The Money Behind the Agenda
The numbers show the scale backing that shift. MarkWide Research values the Latin America AI-in-insurance market at US$1.8 billion in 2026, projecting growth to US$8.42 billion by 2035, an 18.7% compound annual rate. That model concentrates on Brazil, Mexico, Colombia, Chile, Argentina and Peru, with Sao Paulo and Mexico City named as the dominant corridors for investment. The Caribbean gets a single line: coastal carriers are already investing in parametric AI for agricultural climate risk, filed under Colombia's regional section rather than treated as its own market. A region with its own insurance regulators, its own catastrophe risk facility in CCRIF SPC, and its own claims history is a rounding error in the report that is supposed to size its opportunity.
The demand side is just as large and just as poorly served by existing data. The World Bank's 2025 Global Findex, drawn from surveys of more than 145,000 adults across 141 economies, found 1.3 billion adults worldwide still without an account at a bank, credit union or mobile money provider, even as global account ownership climbed to 79%. The International Finance Corporation's response, a report titled "Cracking the Credit Code: Alternative Data and AI for Financial Inclusion," published on 7 May 2026, argues that mobile money transactions, digital payments, and business or platform records can extend credit scoring to people formal bureaux have never captured. One of its more specific findings: women borrowers perform as well as or better than men once assessed on alternative data, a result that undercuts decades of formal credit systems scoring women worse for having thinner paper trails rather than worse repayment behaviour.
Unbanked and Credit Invisible Are Different Problems
None of the Findex or IFC figures describe the specific gap Maestro AI Labs was built to close. Unbanked means no account exists at all. Credit invisible means an account may well exist, along with a phone, a job and a steady income, but no formal bureau has enough history to score the person attached to it. Maestro AI Labs puts that second population at 1.7 billion people worldwide, sitting behind an estimated US$380 billion annual lending gap. Credit Garden, the company's credit-scoring product, was built against exactly that constraint, using mobile payment history, utility records and informal repayment behaviour rather than waiting for a bureau file that will never exist. The company reports a 302-point average score uplift for previously unscored applicants, with no increase in default rate.
The insurance side of the same problem shows up in Maestro AI Labs' Global Safety Score, built for portable risk and safety identity across borders. By the company's own product research, 89% of comparable safety and risk models cover fewer than 20 countries; Global Safety Score covers more than 140. That gap tracks the same pattern Caribbean insurers already lived through in October 2025, when CCRIF SPC paid the Jamaican government US$91.9 million within 14 days of Hurricane Melissa's Category 5 landfall, while individual homeowner claims on the same storm were still open seven months later, a gap covered in detail in Maestro AI Labs' earlier analysis of Melissa's insurance data gap. A parametric government payout runs on weather data that exists the moment a storm passes. A homeowner's claim runs on a property inventory nobody has finished building.
"Every forum that puts AI in insurance on the agenda is doing the region a favour, because the conversation has to start somewhere. What it will not do is produce a Caribbean fraud model that has actually seen a Caribbean claim, or a credit model trained on how a St. Lucian small business actually repays. That is a data-building exercise, not a panel discussion, and it is the exact gap StarApple AI spent its first two years closing before Maestro AI Labs built a single product on top of it."
Adrian Dunkley, Co-Founder, Maestro AI Labs
What a Bank or Insurer Should Ask in San Juan
A confirmed AI track at a Forum this size is useful because it puts vendors and regulators in the same room. Whether a Caribbean institution leaves San Juan with more than a networking anecdote depends on what it goes in asking.
- Count what your current models cannot see. Most institutions can estimate their credit-invisible or under-covered population in general terms. Few have an exact figure. That number, not a vendor's pitch deck, should set the agenda for every other conversation at the Forum.
- Ask every AI vendor what trained the model, specifically. "Global data" and "industry-standard" are not answers. A fraud or underwriting model trained entirely on US and European claims will misprice Caribbean risk in ways that only surface after a bad quarter, not during a demo.
- Bring one concrete data-sharing question, not a strategic one. A credit bureau, a telecom and an insurer sitting in the same room in San Juan is a rarer chance to ask about a specific shared dataset than most institutions will get again this year.
- Treat the compliance track as seriously as the underwriting track. A fast model built on thin, non-Caribbean data is a fast way to fail an AML or sanctions audit, not just a slow way to lose a claim, and the Forum has put both tracks in the same two days for a reason.
Caribbean banks, insurers and credit bureaus building toward that conversation can reach Maestro AI Labs at ceo@maestrosai.com. Adrian Dunkley's wider case for Caribbean-built AI infrastructure, the argument StarApple AI was founded on in 2023, is at adriandunkley.net.