Delivery platforms in Latin America are turning their riders into borrowers, using granular work data to issue credit that is automatically repaid from earnings. A report by Argentina’s central bank shows the number of indebted delivery workers rose 122% in 2025, after a 177% increase the previous year. Average debt for a registered rider reached 900,000 pesos; for merchants borrowing from the same apps, it was seven times higher. PedidosYa, a leading firm, extended 57,000 loans worth $84 million, with monthly payments capped at 30% of the income a worker generates on the platform. The loans are earmarked almost exclusively for buying, maintaining or repairing the motorcycles and bicycles used for deliveries.

The mechanism relies on what the central bank calls ‘digital collateral’. The platform already tracks a rider’s hours, delivery volume, acceptance rate and customer ratings, allowing it to assess creditworthiness without payslips or credit history. Repayment is deducted at source: a portion of each delivery fee is withheld before the money reaches the worker’s account, tying the debt directly to continued platform labour. Labour unions in Argentina argue this creates structural indebtedness. Belén D’Ambrosio of the Sitrarepa union said annualised interest rates can reach 700%, forcing riders to work 10–12 hours a day to cover fixed costs, and is calling for state regulation of these loans.

The same data-as-collateral logic is spreading. In Kenya, the tax authority uses its electronic invoice system to pre-populate returns and cross-check declarations, shifting compliance from self-assessment to data verification. In Brazil, restaurant technology firms offer loyalty and cashback tools that let small eateries capture customer data and reduce reliance on delivery marketplaces, where the platform owns the diner relationship. Both cases illustrate how entities controlling behavioural data are extending financial services to populations without traditional credit histories.

The next milestone is regulatory. Argentina’s central bank has documented the lending surge but has not yet imposed rules; Colombia’s authorities are watching as similar products emerge. Kenya’s revenue authority is pairing its data-driven enforcement with a tax amnesty that expires in December 2026, after which automated assessments are expected to become routine. Whether regulators will introduce interest-rate caps, data-portability requirements or licensing regimes for platform-based credit is the question that will determine how this model evolves.