At Egypt’s Commercial International Bank (CIB), alternative data is not used in isolation but embedded within broader underwriting frameworks. Transactional behaviour – such as utility bills, telecom payments and recurring transfers – provide what the bank describes as a more “dynamic perspective” on a borrower’s financial stability.
“This approach allows CIB to expand access to finance responsibly while strengthening underwriting precision and portfolio resilience,” says Islam Zekry, Group Chief Finance and Operation Officer and Executive Board Member. “By enhancing risk visibility in underbanked segments, CIB extends credit with greater confidence, deeper insight and reinforced quality standards.”
The idea of “responsible expansion” is key to CIB’s strategy. For Zekry, this means improving the quality of data inputs and the discipline around their use in concert. “It is about widening access while continuously strengthening risk calibration and sustaining long-term performance quality,” he says.
Getting this balance right will be critical as lenders across the financial ecosystem tap into new sources of information. Rather than relying solely on periodic financial snapshots, firms are increasingly analysing patterns in real-time or high-frequency transaction activity. For example, Nigeria’s Paga Group uses behavioural signals derived from payment activity to infer reliability and economic stability.
“Traditional credit bureaus measure repayment history — but what they’re really trying to understand is reliability, resilience and economic stability,” says Anthony Isichei, Paga’s General Manager of Consumer Business. “Behavioural transaction data lets us get there from a different angle: not from the rear-view mirror of past debt, but from the real-time texture of someone’s financial life.”
One of the most consistent indicators is the regularity of transactions. Or as Isichei puts it: “The frequency and cadence matter more than the volume.” A street trader processing 20 ₦500 transactions a day tells a different, and often more creditworthy, story than someone who makes one large transfer a month – although this remains context-dependent and must be interpreted alongside income stability and sector risk.
In developed markets, open banking frameworks have accelerated the flow of data across different platforms. In African markets, formal open banking frameworks are nascent, and data-sharing is driven more by fintech platforms and payment ecosystems, as well as partnerships between banks, telcos and fintechs. But this is unlikely to remain the case forever.
“If emerging and frontier markets get to a point where open banking becomes completely integrated, payment flow information from a particular borrower can reveal the [financial] reality much faster,” says Daniel Afolabi, CEO and founder of Nigerian fintech Cede.
The question of accountability
Yet the prospect of faster access to increasing flows of data comes with pitfalls as well as benefits. While it offers lenders a more granular view of financial behaviour, it also introduces new challenges around interpretation and bias. Transaction data, for instance, reflects economic reality. But for poorer borrowers that reality is often the product of deep-seated structural disadvantages. An over-reliance on behavioural data as a proxy for creditworthiness, warns Isichei, could deepen existing inequalities.
The profusion of data flowing between different institutions, technical connections and shared ecosystems also puts accountability in the spotlight. Credit decisions increasingly depend on data sourced from multiple platforms, but it is vital that this does not blur responsibility for errors or misjudgement. “Data contribution can be distributed, but accountability for credit decisions cannot be ambiguous,” says Afolabi.
At present, responsibility is poorly defined across many ecosystems, particularly where data flows across jurisdictions or informal financial networks. As data sources multiply, so too does the burden on lenders to validate, interpret and govern that data.
At CIB, this hasprompted greater emphasis on governance frameworks that can manage both internal models and external data inputs, including strict controls around consent, traceability and model oversight. “Governance is not merely compliance; it is a strategic differentiator,” says Zekry.
The risks associated with alternative data are not limited to governance. As models scale, new forms of behavioural distortion are beginning to emerge. “As behavioural credit scoring becomes more widely understood, customers can optimise their behaviour not to become more creditworthy in reality, but to appear more creditworthy in the model” says Isichei. Such gaming can take multiple forms, from artificially inflating transaction volumes to creating circular payment networks designed to mimic legitimate activity.
The evolution of data-based digital lending brings other systemic challenges. Expanding credit access across different digital platforms without more sophisticated risk assessment could create new cycles of over-indebtedness. For all the talk of constant streams of data, lenders often still lack a comprehensive, real-time view of a borrower’s total exposure across platforms. In such an environment, even responsible individual lending decisions can contribute to unsustainable borrowing in the aggregate.
Yet at the same time, alternative data could help lenders and regulators identify risks earlier. Credit bureau data reflects outcomes that have already crystallised. Transaction data captures behaviour as it evolves. By analysing payment flows and transaction patterns, lenders can detect changes in financial behaviour before they translate into missed repayments or defaults.
Ultimately, alternative data in Africa has the potential to make lending more informed and more inclusive. But none of this is automatic. As information abundance replaces information scarcity, the challenge becomes how to share, govern and interpret a deepening sea of data without introducing new forms of opacity, bias or systemic risk.
