Reviewing the Foundations: Angela Abhulimen and Two Studies on AI and Data in African Finance

By Ugo Aliogo

Two questions have shadowed African financial services for years. Why has insurance failed to win broad public trust, and why has banking struggled to reach the people who most need it? In late 2022, two peer-reviewed review studies took up those questions in detail, and among the researchers behind them is Angela Abhulimen, an independent researcher based in the United Kingdom whose name has become a recurring presence in the literature on African finance and technology.

Both papers appeared in the International Journal of Management and Entrepreneurship Research in its December 2022 issue, and both are the product of a research collaboration rather than a single author. Abhulimen is a co-author on each, working within a group of researchers drawn from finance, insurance, pensions, and payments. For review papers, which set out to survey a whole field, that collaborative model is a strength, because it brings several vantage points to bear on a complicated subject.

The first study, “Artificial Intelligence in African Insurance: A Review of Risk Management and Fraud Prevention,” examines four areas where the technology is starting to bite: underwriting, risk assessment, regulatory compliance, and fraud detection. The authors are precise about the tools involved, from machine learning and predictive analytics to geospatial analysis for flood and drought exposure, and natural language processing and anomaly detection for spotting suspicious claims. They are equally precise about the fraud the sector faces, separating claims fraud, application fraud, premium fraud, and insider fraud, and they ground the analysis in real cases from South Africa, Kenya, and Nigeria, including anti-money-laundering monitoring in Kenya and compliance with South Africa’s Protection of Personal Information Act. Abhulimen is blunt about why penetration stays low. “Low insurance penetration isn’t a mystery,” she says. “It’s the predictable result of weak fraud controls and products built for someone else’s market.”

The second study, “Data Analytics in African Banking: A Review of Opportunities and Challenges for Enhancing Financial Services,” turns to inclusion. Its most consequential argument concerns credit scoring. Conventional scoring relies on formal records that tens of millions of Africans do not have, so the authors examine how alternative data, including mobile-money transactions, airtime and utility payment histories, and mobile-phone usage, can extend credit to people the formal system has never seen. They illustrate the point with banks that have done it, among them Equity Bank in Kenya through its Equitel platform, First National Bank in South Africa, and Standard Bank across several markets. “For African banks, good data isn’t a luxury,” Abhulimen says. “It’s the difference between reaching the unbanked and only talking about them.” The paper is candid that the same techniques carry risks, from data-quality gaps to privacy exposure, and it sets out the conditions, including cloud infrastructure and stronger data governance, under which they pay off.

What gives both reviews their authority is an attention to specifics. Rather than treat African institutions in the abstract, the authors engage with real regulatory environments, infrastructure constraints, skills shortages, and demographic realities, and they name the same recurring obstacles in both sectors: poor data quality, fragmented regulation, thin technical talent, and uneven connectivity. “We weren’t asking whether the technology was impressive,” Abhulimen says of the group’s approach. “We were asking whether it could survive the conditions African insurers actually work in.”

Review papers seldom make headlines, but they tend to set the agenda for the research that follows. These two did precisely that for their authors. They opened a line of inquiry into African finance, technology, and inclusion that Abhulimen and her collaborators would go on to extend substantially over the next two years.