Africa’s financial institutions face a structural reckoning: the tools used to assess creditworthiness are changing faster than the governance frameworks designed to regulate them, and the borrowers of tomorrow bear little resemblance to those the current lending architecture was built to serve.
That was the central tension running through the East African Banking School Conference held recently in Diani, Kenya, where banking professionals, fintech operators and credit specialists examined how artificial intelligence, behavioural analytics, mobile money data and open banking are transforming credit markets across the continent. The implications extend well beyond East Africa. For West African economies navigating sovereign debt constraints, shallow capital markets and persistent financial exclusion, the governance of digital lending has become a first-order policy question.
The numbers frame the stakes clearly. With a median age below 20 years, sub-Saharan Africa’s Generation Z represents the largest untapped credit market on earth. Yet this cohort owns few of the assets that traditional lending models treat as collateral. They rent rather than own homes. They earn income through informal platforms and mobile gig work. They conduct financial lives almost entirely through smartphones. Across ECOWAS member states, where formal credit penetration remains stubbornly low, financial institutions are still largely applying collateral frameworks designed for a borrower profile that no longer dominates the market.
The Data Governance Deficit
Technology has supplied a partial answer. Machine learning models can now process thousands of data points, including mobile money transaction histories, bill payment records, savings patterns and digital footprints, within seconds to estimate default probability. In Kenya, products such as Fuliza and Songesha in Tanzania demonstrate that bank-fintech-mobile operator partnerships can extend credit at scale while building sustainable business models. Ghana’s own mobile money ecosystem, regulated by the Bank of Ghana under its updated Payment Systems and Services Act, has generated comparable transaction data that remains significantly underutilised by formal lenders.
But the conference delivered a pointed institutional warning: algorithms should support lending decisions, not replace professional judgment. Every credit model must be explainable. Where a bank or microfinance institution cannot account for why a customer was declined or approved, it risks embedding systemic bias, degrading governance standards and inviting regulatory sanction. In West Africa, where central banks across the ECOWAS zone are at varying stages of developing AI-specific financial regulation, that accountability gap is real and growing.
Data quality compounds the problem. Inaccurate, incomplete or outdated data produces poor credit decisions regardless of the sophistication of the model applied to it. Financial institutions must build robust data governance infrastructure: information that is reliable, timely, complete and secured against a rising tide of cyber threats. In markets such as Nigeria and Côte d’Ivoire, where credit reference bureau coverage remains partial and data-sharing frameworks between lenders are underdeveloped, the foundational infrastructure for responsible AI-driven lending is still being assembled.
The conference also surfaced a less-discussed credit risk vector: climate change. Following the economic disruption of the COVID-19 pandemic, climate-related shocks, including prolonged Sahelian droughts, coastal flooding in Gulf of Guinea states and erratic rainfall patterns affecting agricultural output across the region, are increasingly threatening household incomes and business continuity. Lenders that fail to integrate climate variables into credit appraisal, portfolio monitoring and stress testing are accumulating risks that their current models do not price.
Responsible Lending as a Governance Standard
Digital lending has expanded financial access. It has also, in several markets, contributed to household over-indebtedness where multiple lenders compete aggressively for the same thin customer base. The conference was direct on this point: responsible lending requires balancing commercial return against customer financial health. That balance is not self-enforcing. It requires regulatory architecture.
For ECOWAS member states, this presents a coordination challenge. National central banks, including the Bank of Ghana, the Central Bank of Nigeria and the Banque Centrale des États de l’Afrique de l’Ouest (BCEAO), which governs the eight-member WAEMU monetary union, each apply distinct consumer credit protection frameworks. The absence of a harmonised regional standard for digital lending creates regulatory arbitrage opportunities, allowing aggressive lenders to operate across borders under the lightest available regime. The ECOWAS Commission’s ongoing work on a regional financial integration framework offers a vehicle for closing that gap, but progress has been incremental.
Conference participants also noted that non-performing loans are frequently created long before borrowers default. Weak credit appraisal, inadequate due diligence and poor underwriting decisions eventually surface as costly recoveries and write-offs. Collections, in this reading, reveal mistakes made at origination. The implication for West African institutions carrying elevated NPL ratios, Ghana’s banking sector NPL rate stood above 20% in 2023 according to Bank of Ghana data, is that the solution lies upstream in credit assessment quality, not downstream in debt recovery intensity.
Progressive lending models offer one structural response. Smaller initial loan facilities, scaled upward as borrowers demonstrate repayment discipline, limit downside exposure while building the transaction histories that improve future credit decisions. Behavioural consistency, cash-flow regularity and demonstrated financial discipline can function as collateral substitutes for borrowers who hold none of the assets that traditional security frameworks recognise.
The institutions best positioned to capture West Africa’s next generation of borrowers will be those that resolve this governance equation: combining high-quality data infrastructure, explainable AI models, skilled credit professionals and regulatory compliance into a lending architecture that is both commercially viable and socially responsible. For investors assessing financial sector exposure across the region, the quality of an institution’s data governance and credit culture will increasingly determine its long-term portfolio health. That is not a technology question. It is an institutional one.
