“Data is the collateral”: Sparkle CEO Uzoma Dozie on fixing Nigeria’s SME lending blind spot
Uzoma Dozie, CEO of Sparkle MFB, says Nigerian banks must rethink collateral requirements and use transaction data and AI to expand lending to small businesses, lower borrowing costs, and improve financial inclusion.
Former Diamond Bank chief Uzoma Dozie, now CEO of Sparkle MFB, argues that African banking is trapped in a collateral trap and that digital banks must move beyond one-to-one payments. He believes AI can reduce the cost of SME lending. Across Nigeria's banking industry, small enterprises and mass-affluent individuals account for a large portion of the deposit base that funds commercial balance sheets, yet receive only 1% of total bank lending.
Sparkle CEO Uzoma Dozie claims that conventional lenders are hindered by an outdated obsession with physical collateral, which prevents them from lending to small businesses and individuals who lack sufficient information for banks to trust. To address this, Dozie suggests redefining collateral using transactional data, which can provide insights for credit decisions.
He regrets the industry's rush to adopt superficial digital labels and emphasizes that technology in banking is an extension of second-generation reforms that improved customer convenience. Sparkle, a microfinance bank, targets mass-affluent individuals and small enterprises, providing business support services and capturing business data to make better decisions and help businesses access credit.
Dozie also challenges the way financial inclusion is measured, arguing that simply tallying the number of bank accounts ignores the need for structural integration, such as linking government social support programs through formal banking rails to give citizens a reason to use an account. For Sparkle, meaningful inclusion means simplifying operational cash flows for small businesses and households, including payroll, inventory management, and paying domestic staff.
Written by urgent.news from TechCabal's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.