Nedbank taps AI-powered lending to reach underserved South Africans

Nedbank, one of South Africa’s largest banks, is turning to artificial intelligence to expand lending to millions of customers who have traditionally fallen outside the reach of formal credit.

The Johannesburg-based lender has partnered with JUMO, a fintech infrastructure company, to launch Nedbank Quick Loans, an AI-powered lending product embedded within the Nedbank Money App. Customers can apply for loans from as little as R500 ($27) and receive credit decisions within minutes, with repayment terms ranging from one to 12 months.

The partnership indicates a broader shift in African banking, with lenders increasingly turning to fintechs and AI to reach customers that conventional credit models have struggled to serve. 

The real test will not be how much money the partnership lends, but whether it can profitably offer fair, transparent credit to millions of low-income and thin-file borrowers who have historically relied on loan sharks, payday lenders and other informal lenders.

“By embedding JUMO’s AI-led proprietary models and processes into Nedbank’s Money App, we are enabling a quick and easy client experience,” said Mutsa Chironga, Managing Executive for Personal Banking at Nedbank. “It is a great solution for smaller and short-term borrowing needs clients have.”

Nedbank, which has a market capitalisation of R127 billion ($6.9 billion), is providing the balance sheet and banking infrastructure. JUMO, which operates in eight African markets and has facilitated more than $10 billion in loans across the continent, supplies the AI-powered lending engine that evaluates borrowers in real time.

“We’ve spent the last 11 years building banking infrastructure and earning the trust of banks,” JUMO CEO Paul Whelpton told TechCabal on Thursday. “What we’re now seeing is JUMO becoming the intelligence layer of African financial services.”

Historically, small-ticket lending has been difficult for banks. Customer onboarding costs were high, risk assessment was expensive, and predicting defaults among borrowers with limited credit histories was challenging. As a result, 37% of consumers who turned to the short-term informal credit market, dominated by more than 40,000 loan sharks, eventually required debt counselling.

JUMO believes AI can change that equation.

According to Whelpton, the platform analyses transactional, behavioural and repayment signals to assess both a customer’s ability and willingness to repay a loan.

“We use behavioural and transactional information, as well as signals we have learned over the last decade about how customers behave outside the traditional banking and credit world,” he said.

Whelpton noted that JUMO’s technology can help lenders serve customers who lack formal payslips, audited financial statements or extensive credit bureau histories.

“I think the average man on the street, or the grandmother in the township, will now get access to formal credit in the right way,” he said. “It’s fully transparent, with no hidden fees and nothing that could hurt them or get them into a debt trap.”

Questions remain about whether easier access to digital credit could create new forms of over-indebtedness. Whelpton said JUMO and Nedbank have implemented strict affordability measures and lend below regulatory maximum thresholds to avoid pushing customers to their limits.

The economics appear promising. JUMO says it has disbursed more than $10 billion across Africa and processes about $240 million in loans each month. Despite focusing on unsecured lending and customers with limited traditional credit histories, its banking partnerships are recording default rates of about 3.3%.

The partnership offers a glimpse into South Africa’s banking future, where banks provide the capital and trust, while fintechs deliver the intelligence that determines who gains access to credit.