Sparkle CEO: African banking stuck in collateral trap, data and AI can unlock SME lending

At AFIS 2026 in Luanda this week, Sparkle MFB CEO Uzoma Dozie argued that Nigerian banks are trapped in an outdated collateral model: small businesses provide the bulk of deposits but receive less than 1% of total lending because banks demand physical assets they don't have. Dozie, formerly Diamond Bank CEO, said the fix is redefining collateral as transactional data — every sale, payment, and cash flow becomes the credit signal. "If we see data as information that generates insights, we can use it to make credit decisions," he said, noting this can be less risky than traditional collateral.

Sparkle, licensed as a microfinance bank, skips the human-touch model entirely — no account officers — to serve mass-affluent individuals and small enterprises at scale. Its differentiation isn't peer-to-peer transfers ("anyone can do that") but one-to-many flows: payroll, batch payments, supplier disbursements. These transactions generate the business data Sparkle uses to underwrite loans and connect merchants with investors.

On AI, Dozie dismissed hype: conversational AI handles balance checks and support, stripping admin costs; machine learning analyzes cash flow patterns against market dynamics to build financial profiles paper statements can't. This lowers lending costs and could mean cheaper rates for SMEs. Merchants also get operational insights from their own data. Dozie warned that financial inclusion isn't counting bank accounts — with 100 million Nigerians below the poverty line, accounts without economic activity are "ornamental." Real inclusion means channeling government social support through formal rails and simplifying business cash flows.

As digital banks turn your transaction history into credit scores, would you move your business banking to a platform that lends based on cash flow rather than land documents?

SOURCE: https://techcabal.com/2026/11/05/uzoma-dozie-ai-sme-lending-nigeria/