Speaking on the sidelines of AFIS 2026 in Luanda, Angola, Uzoma Dozie, the former Diamond Bank chief andSparkle MFB CEO, argues that African banking is stuck in a collateral trap, explains why digital banks must move beyond one-to-one payments, and details how AI can reduce the cost of SME lending. GODFREY GEORGE reports.
Across Nigeria’s banking industry, small enterprises and mass-affluent individuals account for a substantial share of the deposit base that funds commercial balance sheets. Yet, when those same small businesses apply for credit, they receive less than 1% of total bank lending.
For Uzoma Dozie, Managing Director and Chief Executive Officer of Sparkle Microfinance Bank, this disparity captures why capital is failing to support the real economy.
Speaking to TechCabal at the Africa Financial Industry Summit (AFIS) in Luanda, where the central theme was “Making Capital Count”, the veteran banker argued that conventional lenders are held back by an outdated obsession with physical collateral, locking out the very businesses driving national employment and gross domestic product.
“Capital tends to flow towards areas of certainty, even though every investment carries some level of risk,” Dozie said. “What drives capital, particularly from a banking perspective, is trust. Take big businesses and large corporations: their accounts are audited and verified by professionals, so banks can lend based on information they trust.”
“But when you move down to small businesses and individuals, you are looking at the person. You may understand that they have a business, but you do not have sufficient information about that business. The bank either decides not to lend or lends at a price that reflects the perceived risk.”
To break this impasse, Dozie argues that the definition of collateral must be completely overhauled through transactional data.
“Every small business generates data. Every transaction is data,” Dozie said. “If we see data as information that can generate insights, we can use it to make credit decisions. For me, that is part of making capital count: changing our mindset about what constitutes collateral for small businesses and scaling lending by examining cash flows and business data. In some cases, this could even be less risky than traditional collateral-based lending.”
Moving beyond buzzwords
Dozie, who led Diamond Bank before its merger with Access Bank, rejects the industry’s rush to adopt superficial digital labels.
To him, technology in banking is not a recent novelty but a continuation of second-generation banking reforms that introduced automated teller machines (ATMs), Internet portals, and mobile banking applications to improve customer convenience.
Licenced as a microfinance bank, Sparkle, he noted, was designed to apply those lessons to segments that traditional retail banking fails to serve efficiently.
“When I hear people talk about wanting to become a digital bank, I think we have created names and buzzwords without really looking beneath the surface,” Dozie observed.
“The difference between our service model and traditional personal banking is that we do not have people involved in every customer interaction. We do not rely on account officers to deliver our services. You cannot bank millions of Nigerians using a purely human-touch model. There is a limit to the consistency, repeatability, and operational efficiency you can achieve that way.”
Rather than competing directly for the unbanked mass market or chasing standard retail transactions, Sparkle targets mass-affluent individuals and small enterprises whose economic activities intersect with the broader population.
“An affluent individual in Nigeria probably has four or five people they pay, including people who run small businesses,” Dozie explained. “On the small-business side, it was never just about providing a bank account. A bank account is a given because every business needs one. What was important to us was providing business support services.
“We built these services to capture business data. Once we capture that data, we can use it to generate intelligence about the business, make better decisions, and help the business access credit or even connect with people interested in investing in it.”
Rethinking financial inclusion
Dozie also challenged the way the financial sector evaluates financial inclusion, cautioning that tallying the number of registered bank accounts can create a misleading impression of progress.
In an economy where an estimated 100 million people live below the poverty line, an account without an economic foundation remains largely ornamental.
“If we do not solve the problems of identity and social inclusion, financial inclusion will not happen,” Dozie said. “A bank account is useful for making and receiving payments, as well as storing money. But if you have 100 million people living below the poverty line, financial inclusion cannot simply be about opening bank accounts for them.”
Meaningful inclusion, Dozie argued, requires structural integration where government social support—ranging from healthcare and insurance to emergency relief and education—is channelled directly through formal banking rails, giving citizens a clear reason to use an account.
At the enterprise level, inclusion hinges on simplifying operational cash flows. Small businesses must pay staff, manage inventory, and disburse wages, while households must pay domestic staff like security guards and gardeners.
This operational reality has shaped Sparkle’s transaction focus. While many digital banks concentrate on individual, peer-to-peer (one-to-one) transfers, Sparkle has prioritised collective and multi-party flows: one-to-many payments, payroll schedules, and batch transfers.
“Anyone can facilitate one-to-one transactions,” Dozie noted. “But one-to-many, collective payments, and many-to-one transactions are areas where Sparkle specialises. If you look at what we offer, you will see that it is quite different from what other players are doing.”
Conversational services and cheaper credit
On artificial intelligence, Dozie dismissed the notion that the technology is merely speculative, pointing to practical deployments across Sparkle’s marketing, customer engagement, and credit operations.
At the interface level, conversational AI allows users to communicate in their preferred format and check account balances directly, stripping out administrative costs that previously required separate software tools or customer service staff.
For commercial borrowers, machine learning algorithms analyse internal cash flows, matching payment patterns against market dynamics to build a more accurate financial profile than traditional paper statements permit.
“Internally, AI allows us to use customers’ cash flows to identify insights and correlations between their payments, their markets, and their business activities,” Dozie said. “We can then use that information to make better credit decisions. This helps us reduce the cost of lending and potentially offer credit at lower interest rates than would ordinarily be possible through traditional lending models.”
Dozie added that this analytical capability is also available to merchants, providing them with insights into their operational performance and customer behaviour.
“They give us access to their data, and we use it to provide insights that help them make better decisions,” Dozie said. “At the same time, we use that information to make better decisions about how we serve them and grow alongside their businesses. For us, it is not just a buzzword. We are only scratching the surface.”
