Fix it before you speed it up: choosing what to automate first

Small businesses often automate the shop window while leaving the stock room, payroll and ledger to manual work. Across the Caribbean, only 4 to 5 per cent of business transactions are conducted digitally, compared with 18 to 19 per cent in the United States, according to figures cited by Barbados Senior Minister Kerrie Symmonds.

The imbalance is especially clear among Barbadian micro, small and medium enterprises. Seventy-eight per cent use social media and customer-facing platforms, but only 13 per cent use payroll software and 14 per cent use digitised inventory management. One quarter use no digital management tools at all. Jamaica shows a similar pattern. A Mastercard study found that just 8 per cent of small merchants use point-of-sale systems, while cash accounts for about 72 per cent of personal consumption spending. Jamaica has roughly 422,000 registered micro, small and medium enterprises.

The same order appears in the United States. An Intuit-commissioned survey of more than 2,200 businesses with up to 100 employees found that marketing and customer service were the most common uses of artificial intelligence, at 43 per cent and 36 per cent respectively. Bookkeeping ranked last, at 29 per cent. Yet finance functions often offer the clearest opportunities for automation. The Hackett Group’s 2026 study found accounts payable to be the most mature area of AI adoption, with one-third of organizations already scaling solutions there, followed by travel and expense management.

The difference comes down to the nature of the work. Supplier payments and expense processing are high-volume, repetitive and governed by stable rules. Forecasting is less predictable, while marketing copy may feel satisfying to automate without changing the cost base much.

The practical starting point is therefore not the tool itself. Businesses should first ask how often a process occurs, whether its rules remain stable, whether the inputs are consistent and whether they would notice if something went wrong.