By KABIR FAGGE

As banks across Nigeria and West Africa race to modernise, a quiet transformation is taking place. Not in front-facing apps or glossy fintech interfaces, but deep within their data systems. At the core of this shift is a growing reliance on applied machine learning to improve customer targeting, reduce risk, and guide branch and product decisions. 

One contributor to this evolution is a Nigerian data scientist. Over the last five years, he has worked behind the scenes on analytics systems used by several banks and digital platforms operating across Africa. His name doesn’t appear in press releases, but teams inside the institutions credit his models with improving how decisions get made, from which customers receive credit offers to how relationship managers prioritise leads. This change maker is no other than Adam Usman.  

The graduate of Ahmadu Bello University, Zaria and Northeastern University, Boston, USA, began his work in financial services during his time at Bluechip Technologies, a Nigerian data consultancy that supports tier-one banks and telecoms. There, he helped implement credit scoring frameworks that went beyond traditional bureau data, using mobile transaction behaviour and income volatility indicators to flag potential defaults. One project, run in partnership with a major Nigerian bank, reduced early delinquency rates by nearly 20 per cent among digital loan applicants. The improvement was confirmed in an internal audit seen by the bank’s credit risk team. 

Beyond lending, Usman also supported product and marketing teams with targeting models designed to improve digital cross-sell. These models segmented customers not just by balance and tenure, but by real-time behaviour, including frequency of POS transactions, dormant account signals and customer service interaction patterns. The result was a shift from batch-style campaigns to tailored product recommendations that could be deployed via email, SMS and in-app banners. 

Colleagues say the value of the models wasn’t just in their complexity, but in how they were designed to fit into existing workflows. “It was never just data for the sake of it,” said a product lead at a commercial bank who worked with Usman on segmentation logic. “His strength was turning the technical stuff into tools people could actually use.” 

After moving to the U.S., Usman joined Toptal, a global remote talent marketplace, where he shifted his focus to growth analytics and SEO modeling. At Toptal, he designed a system that mapped how internal link structure and external backlinks influenced search visibility, borrowing concepts from Google’s PageRank to simulate authority flow across hundreds of landing pages and articles. The project helped inform how content was prioritised and rewritten and internal metrics shared with the executive team showed a 30 per cent lift in SEO-attributed traffic within one quarter of deployment. 

Usman also led efforts to consolidate marketing and operations data into a certified repository, enabling faster decision-making across the growth and finance teams. This internal platform became the backbone for several A/B tests and customer retention experiments, which were previously stalled due to fragmented and conflicting metrics. 

His recent work, shared through API libraries and open-source scoring templates, has begun to gain attention in early-stage fintech ecosystems across West Africa. One accelerator in Accra now uses a stripped-down version of Usman’s credit model to train startups in responsible lending design. In a remote training session earlier this year, attended by developers and analysts from six countries, he walked participants through how to build scoring features using customer interaction logs and fraud-risk heuristics. 

The methods are gaining interest at the policy level as well. Sources at the Central Bank of Nigeria say the institution is reviewing proposals for explainable credit scoring systems and that pilot efforts are underway to compare traditional bureau scores with models that include behavioural signals. Usman’s work has been cited in at least one internal benchmarking document related to model governance and fairness auditing. 

While his name remains little-known outside of data and banking circles, the models he helped build continue to shape how financial institutions in Africa make decisions. From loan approvals to how they allocate relationship managers and evaluate branch productivity. Quietly, this kind of work is changing the way banking is done in emerging markets. Less through flashy apps and more through careful, structured data decisions behind the curtain.•Fagge, an IT expert and public affairs analyst, wrote from Abuja