Access to credit is a fundamental enabler of economic opportunity, yet it remains out of reach for more than one billion adults around the world.
Written by Francois Grobler, Head of Decision Analysis, Experian Africa
In South Africa, the challenges are particularly acute. More than 85% of small businesses seeking financing have a turnover of less than R1 million and have the highest rejection rates from traditional credit scoring models.
This reality highlights an important question: How can lenders responsibly expand access to credit without compromising risk integrity? The answer may lie in machine learning (ML) and alternative data.
A recent study commissioned by Experian and conducted by Forrester Consulting found a growing consensus among senior decision makers about ML’s potential to democratize trust. 70% of respondents across 11 countries believe that improved ML accuracy will enable them to serve consumers who have been denied credit, an important step toward a more inclusive financial ecosystem.
This article explores how these technologies can transform credit decisions in South Africa and create opportunities for underserved consumers and micro, small and medium enterprises (MSMEs).
Why machine learning is important for financial inclusion
Machine learning models are revolutionizing risk assessment. ML allows for more accurate predictions of repayment behavior by analyzing vast data sets and identifying patterns beyond traditional scorecards.
The results in South Africa are convincing. 93% of surveyed organizations using ML reported improved credit card approval rates, and 89% saw a reduction in bad debts. For lenders, this means smarter decisions and the ability to offer services that benefit new markets. For consumers, that means opening the door to financial inclusion.
The research results are clear. ML doesn’t just improve risk models. It is redefining inclusion in a diverse and rapidly evolving economy like South Africa.
Alternative data: the missing piece in South Africa’s credit puzzle
Alternative data such as utility payments, rental history, and mobile transactions are essential for evaluating thinner customers with limited traditional credit history.
In South Africa, this group includes young people, gig workers and many MSMEs operating in the informal cash economy.
Ignoring this segment means leaving a large part of the population out of formal credit.
For these economically active consumers, many of whom are small business owners, alternative data provides valuable insight into their spending habits and ability to repay. More than three-quarters (77%) of credit risk decision makers surveyed agree that alternative data is key to improving lending accuracy.
Combining this data with ML enhances decision-making, with 71% of respondents saying that reliably evaluating thinfile customers improves profitability.
Real-world applications drive change
Initiatives like open banking, which allow consumers to securely share transaction data, are gaining traction around the world. This data provides detailed insight into your income and expenses, helping lenders make fairer decisions.
Across EMEA, 86% of businesses have invested or plan to invest in open banking, and more than half are already seeing significant value. For lenders, early adoption provides a competitive advantage. For consumers, that means faster and fairer access to credit.
The question is not if Open banking will transform lending in South Africa, but how quickly The people will accept it.
The promise of ML for small businesses
MSMEs are the backbone of South Africa’s economy, contributing over 40% of GDP and employing over 60% of the workforce. But their growth is often constrained by manual and paper-heavy credit evaluations.
ML is trying to change this. South African organizations using ML report significantly improved SME loan acceptance rates, contributing to business growth and job creation.
This trend is mirrored across EMEA and Asia Pacific, with 88% of companies surveyed seeing similar improvements, highlighting the technology’s potential to create economic opportunities.
The future of financial decision making
The momentum of ML is undeniable. Nearly 8 in 10 organizations already using ML in South Africa plan to significantly increase their investment over the next 1-3 years.
However, barriers such as cost, lack of understanding, and legacy IT infrastructure continue to slow progress for non-adopters.
As financial institutions embrace AI and ML, the question is no longer just “how fast?” But “for whom?”
South Africa’s experience shows that inclusion and innovation can go hand in hand. Expanding access to credit fosters entrepreneurship, job creation and economic resilience. In a market where financial exclusion has long been a barrier, this technological shift could be a turning point towards inclusive prosperity and sustainable economic transformation.
