On July 12, 2026, the Planning Institute of Jamaica confirmed what the diaspora already knew from their own bank statements: net remittances into Jamaica reached US$3,247.5 million in 2025, up 3.8 percent on 2024. Bank of Jamaica figures put the gross number higher still, at US$3.49 billion in against US$238.2 million out. The United States remained the largest single source at US$2,103.1 million, up 5.4 percent. The United Kingdom sent US$352.0 million, up 10.5 percent. Canada sent US$295.1 million, up 3.5 percent.
Those figures describe something bigger than a line item in an economic survey. Remittances were worth roughly 15.3 percent of Jamaica's GDP in 2024, according to Bank of Jamaica and World Bank data, more than three times the global average of about 5 percent. It is money that pays school fees in Mandeville, keeps a grandmother's medication stocked in Portmore, and covers the gap between a monthly salary and a monthly bill for hundreds of thousands of Jamaican households that a spreadsheet in Kingston will never see.
I have sat across the table from Caribbean bank executives more than a hundred times over the past two years, training boards on AI adoption through StarApple AI. The subject that comes up in nearly every one of those rooms, ahead of fraud, ahead of digital transformation strategy, is correspondent banking: how much longer the correspondent account survives, and what happens to a remittance-dependent economy the day it does not.
The Number Behind the Record
The PIOJ's 2025 Economic and Social Survey lands at a moment when Jamaica needed the good news. Hurricane Melissa tore through the west of the island in late 2025, and the Bank of Jamaica has already pushed its recovery timeline out to three or four years. Remittances did what they tend to do during a Caribbean recovery: they rose. Diaspora giving typically climbs after a disaster, as family abroad send more, not less, when things at home get harder. The 3.8 percent increase sits on top of steady wage growth among Jamaican workers in the United States, the United Kingdom, and Canada.
The concentration matters as much as the total. Two of every three Jamaican remittance dollars arrive through a single corridor from the United States. Add the United Kingdom and Canada and three countries account for the overwhelming majority of the flow. That is not a diversified income stream. It is a small number of banking relationships, correspondent accounts, and money transfer operator licences standing between a Jamaican household and its monthly income. Every one of those relationships depends on a foreign bank deciding that the compliance cost of serving a Caribbean correspondent is worth carrying.
The Pipes Carrying That Money Are Narrowing
That decision has been going the wrong way for over a decade. The Bank for International Settlements has tracked a 25 percent decline in active correspondent banking relationships worldwide since 2011, with the steepest cuts landing in Africa, the Pacific Islands, and the Caribbean. A 2017 survey by the Caribbean Association of Banks found that 21 of its 23 member countries had lost at least one correspondent banking relationship. Using SWIFT data from 2011 to 2020, researchers tracking the region found that Belize, Saint Vincent and the Grenadines, Dominica, the Bahamas, and Jamaica each lost more than 40 percent of their correspondent banking counterparties. The World Bank has named the Caribbean the region most severely affected by de-risking anywhere in the world, according to Atlantic Council research on the issue.
This is not an abstract regulatory story. When a correspondent bank exits a Caribbean market, the local banks and money transfer operators that relied on it either find a replacement, usually at a higher price, or lose the ability to move US dollars in and out of the country. Smaller operators, the ones most likely to serve rural parishes and lower-income households, are the first to be cut off. The families on the receiving end pay for that exit twice: once in higher transfer fees, and again in the risk that a channel disappears and pushes them onto something slower or less regulated.
Why Global Banks Keep Walking Away
The banks leaving are not leaving because the Caribbean is unprofitable. Correspondent banking with a small island jurisdiction was rarely a major revenue line for a global institution in the first place. Banks are leaving because the compliance cost of proving, to their own regulators, that every Caribbean correspondent is not a money laundering or sanctions risk has become expensive enough that the revenue no longer covers it. A string of multibillion-dollar anti-money laundering enforcement actions against major global banks over the past fifteen years taught the industry a hard lesson: the cost of a compliance failure vastly exceeds the profit from a low-volume correspondent account.
The result is a blunt calculation that has little to do with whether a specific Caribbean bank is well run. Global compliance teams increasingly manage risk by geography rather than by institution, treating an entire region as a single risk category and cutting relationships in bulk rather than assessing them individually. A well-governed bank in Kingston or Bridgetown gets caught in the same net as a poorly supervised one three islands away, because proving the difference, one correspondent at a time, costs more than most global banks are willing to spend on a market this size.
AI Compliance: The Case Caribbean Banks Need to Make
This is precisely the problem AI-driven compliance systems are built to solve, and the industry evidence is now substantial. The UK's Financial Conduct Authority reports that 75 percent of regulated firms already use AI in some part of their compliance function, with a further 10 percent planning adoption within three years. Machine learning transaction monitoring systems can process the full volume of a bank's payment traffic in real time, flagging genuinely suspicious activity while cutting the false positive rate that currently forces compliance teams to review thousands of harmless transactions for every one that matters.
For a Caribbean bank trying to keep or rebuild a correspondent relationship, that shift changes the conversation with a global partner. Instead of asking a foreign compliance officer to take supervision quality on faith, a Caribbean bank running AI-powered transaction monitoring, sanctions screening, and beneficial ownership verification can produce the kind of continuous, auditable evidence that global risk committees are actually asking for. Anti-money laundering technology firm Sumsub reported in 2026 that institutions using AI for fraud and compliance detection cut fraud losses by 40 percent while sharply reducing manual review volume. That is the argument a Caribbean bank needs walking into a correspondent renewal meeting: not a promise of good behaviour, but a system generating the audit trail a global bank's own regulator will accept.
None of this reverses de-risking on its own. A correspondent bank that has already decided the Caribbean is not worth the trouble will not be won back by a dashboard. But AI compliance tooling does something more achievable: it lowers the cost of serving a Caribbean correspondent well enough that staying becomes the easier decision, and it gives regional banking associations an evidentiary case to bring to regulators in Washington, London, and Ottawa when they argue that blanket regional de-risking is disproportionate to the risk any single institution presents.
Cheaper Transfers: What AI-Native Payment Rails Change
The correspondent banking question sits upstream of a second problem families feel directly: the cost of sending the money in the first place. The Caribbean fintech market, valued at roughly US$1.2 billion in 2025, is projected to reach US$2.8 billion by 2030, according to research firm Hope Research Group, and remittances are the largest single use case driving that growth. Fintech platforms that settle payments through local account networks rather than routing every transaction through a chain of SWIFT intermediaries can move money faster and at lower cost, because they sidestep some of the correspondent relationships that are shrinking in the first place.
AI does the specific work inside those platforms that makes lower pricing possible without lowering safety. Machine learning fraud models screen transactions in milliseconds instead of the manual holds that slow down traditional wire transfers. AI-driven treasury tools optimise which currency corridor and settlement route a transfer uses in real time, the way a mapping app routes around traffic. For a Jamaican family receiving a transfer every month, the gap between an AI-optimised fintech corridor and a legacy wire service is not marginal. It can be the difference between losing a small share of the transfer to fees and losing considerably more, money that would otherwise cover a week of groceries or a light bill.
The Unbanked Opportunity Nobody Is Pricing In
There is a second opportunity layered underneath the first, and it gets less attention than it deserves. In rural parishes and smaller Caribbean territories, cash pickup remains the primary way many recipients receive remittances, because branch banking access is limited or the paperwork required to open an account is out of reach. That group has almost no credit history in any formal sense, which shuts them out of loans and business financing even when they receive a stable, predictable monthly income from abroad.
AI changes what counts as a credit signal. A recipient who has received a remittance transfer of a similar size on a similar schedule for two years has a financial track record every bit as real as a pay stub, even though no traditional lender reads it that way. Machine learning underwriting models can build a credit profile from remittance receipt patterns, utility payment history, and mobile money transaction data, extending small loans and formal financial products to households that have been creditworthy in practice for years and invisible to a bank on paper. Development finance institutions have piloted versions of this approach elsewhere in the world, with real repayment data behind the results.
This is not a hypothetical add-on. It is the more durable answer to the same underlying vulnerability: an economy where hundreds of thousands of households depend on a financial channel they do not own, cannot influence, and have no formal relationship with beyond the moment the cash changes hands. AI-based financial inclusion tools turn that one-way dependency into an actual banking relationship.
What Has to Happen Next
Three things, and none of them require new institutions.
Caribbean banking associations should commission a joint AI compliance pilot. The Caribbean Association of Banks, which has tracked de-risking since 2017, is the natural body to run a shared transaction-monitoring and sanctions-screening platform that smaller member banks could not afford individually but can access collectively, generating the audit trail correspondent partners are asking for at a fraction of the per-institution cost.
Regulators across CARICOM should formally recognise AI-verified compliance evidence in their supervisory frameworks. If a Caribbean central bank certifies that a commercial bank's AI monitoring system meets an agreed standard, that certification becomes something a correspondent bank's compliance team can cite directly, shortening the due diligence cycle that currently makes every Caribbean correspondent relationship a multi-month negotiation.
Caribbean fintech builders should treat the unbanked remittance recipient, not the diaspora sender, as the real product opportunity. The sending side is already served by Western Union, MoneyGram, Wise, and a growing list of competitors. The receiving side, an AI-underwritten financial product built around a Caribbean household's actual remittance history, is close to unclaimed. Barbados has already positioned itself as a fintech testing ground, and CARICOM tax administrators who met in Georgetown, Guyana on July 15, 2026 are part of the same regional push toward stronger financial data infrastructure. The pieces exist. They have not yet been assembled around the people who need them.
None of this changes the basic fact that a Jamaican family's monthly income depends on decisions made in compliance departments the family will never see, in cities most of them will never visit. AI does not remove that dependency. What it can do is make the Caribbean side of the relationship cheaper to serve, safer to certify, and harder to walk away from, while opening a second front, financial inclusion for the people receiving the money, that nobody has built yet at scale.
US$3.25 billion crossed into Jamaica in 2025 through banking pipes that keep getting narrower. The PIOJ's report is a milestone worth marking. It is also, if the correspondent banking trend lines hold, a number that gets harder to repeat every year the region does not act on what is shrinking underneath it.
Frequently Asked Questions
Why did Jamaica's remittances hit a record in 2025?
Net remittances into Jamaica reached US$3,247.5 million in 2025, up 3.8 percent on 2024, according to the Planning Institute of Jamaica's 2025 Economic and Social Survey, reported on July 12, 2026. Bank of Jamaica data puts gross inflows at US$3.49 billion against outflows of US$238.2 million. The rise reflects steady wage growth among Jamaican workers abroad and the pattern, common after a major disaster, of diaspora families sending more money home. Hurricane Melissa struck western Jamaica in late 2025, and remittance flows typically increase during Caribbean recovery periods as family abroad step up support.
What is correspondent banking de-risking?
Correspondent banking de-risking is the practice of a global bank ending its banking relationship with a smaller foreign bank, often because the cost of proving that relationship meets anti-money laundering and sanctions compliance standards outweighs the revenue it generates. The Bank for International Settlements has tracked a 25 percent decline in active correspondent banking relationships worldwide since 2011, with the steepest losses in Africa, the Pacific Islands, and the Caribbean. A 2017 survey by the Caribbean Association of Banks found that 21 of its 23 member countries had lost at least one correspondent relationship.
Which Caribbean countries are most exposed to correspondent banking de-risking?
Using SWIFT data from 2011 to 2020, researchers found that Belize, Saint Vincent and the Grenadines, Dominica, the Bahamas, and Jamaica each lost more than 40 percent of their correspondent banking counterparties over that period. The World Bank has identified the Caribbean as the region most severely affected by de-risking anywhere in the world. Smaller territories with lower transaction volumes tend to be the most exposed, because they offer global banks the least revenue to offset compliance costs.
How can AI help Caribbean banks keep their correspondent banking relationships?
AI-powered transaction monitoring, sanctions screening, and beneficial ownership verification give Caribbean banks a continuous, auditable compliance record they can present directly to correspondent partners and regulators. The UK's Financial Conduct Authority reports that 75 percent of regulated firms already use AI in some part of their compliance function. Rather than asking a foreign bank to take supervision quality on faith, a Caribbean bank running AI compliance tools can demonstrate it in real time, lowering the cost of serving that relationship well enough to make staying the easier decision.
Can AI actually lower the cost of sending money to the Caribbean?
Yes, mainly through fintech platforms that use AI to route transfers through the cheapest available currency corridor and settle payments locally rather than through a chain of correspondent banks. The Caribbean fintech market was valued at roughly US$1.2 billion in 2025 and is projected to reach US$2.8 billion by 2030, according to Hope Research Group, with remittances as the largest driver of that growth. AI fraud screening also replaces the manual holds that slow down and add cost to traditional wire transfers.
How does AI help unbanked remittance recipients build credit?
Machine learning underwriting models can treat a consistent history of remittance receipts, alongside utility payment and mobile money data, as a credit signal, even for recipients who have never held a formal bank account. A household that has received a similar-sized transfer on a similar schedule for two years has a real financial track record that traditional lenders are not set up to read. AI-based underwriting converts that pattern into eligibility for small loans and formal financial products.
Is AI compliance technology affordable for small Caribbean banks?
The most viable path is a shared regional platform rather than individual bank purchases. A body such as the Caribbean Association of Banks, which has tracked de-risking since 2017, could commission a joint AI transaction-monitoring and sanctions-screening system that smaller member banks access collectively, spreading the cost across the region while still producing the individualised audit trail each bank needs for its own correspondent relationships.
What should Caribbean governments and regulators do about de-risking right now?
Three actions matter most: fund a shared AI compliance platform through regional banking associations, have CARICOM central banks formally recognise AI-verified compliance evidence in their supervisory standards so correspondent banks can rely on it directly, and direct fintech investment toward AI-underwritten financial products for remittance recipients, not just cheaper transfers for senders. All three build on infrastructure and institutions that already exist.
"The families sending that US$3.25 billion did their part. The compliance systems standing between them and the people receiving it have not caught up. AI is how the Caribbean catches up before the next correspondent bank walks away." - Adrian Dunkley, the AI Boss