Machine Learning Helps Expand Credit Access

Artificial intelligence (AI) is helping increase access to financial services in Africa.

In recent years, advances in machine learning, a type of AI, have had a profound effect on the delivery of financial services, helping to democratize access in Africa’s emerging economies.

For example, it is being used to offer loans and credit opportunities to people who might otherwise be excluded from the financial system.

AI companies such as the Dubai-based FinTech Optasia are using machine learning in their credit decision engines to automatically approve applications for microloans, helping to expand access to credit.

While not a lender itself, Optasia’s technology is integrated into the lending process, enabling banks and other FinTechs to assess the risk of non-payment automatically, leading to faster decisions and more accessible lending products.

In one recent partnership, Optasia has teamed up with Ecobank and MTN to offer micro-loans to MTN’s customers in Guinea. With capital provided by Ecobank and disbursement handled by MTN mobile money, Optasia’s AI platform provides the crucial risk assessment that facilitates the loans.

Machine learning also allows lenders to deploy more diverse datasets in their decision-making processes. Unlike traditional credit scoring methodologies that require electronic transaction data to build a credit file, a generation of African innovators like Optasia are leveraging alternative datasets to prove the likelihood a given borrower will default on their payments.

And because telecom companies like MTN have access to a wealth of data on African consumers, they have been at the forefront of innovation in alternative credit scoring.

Still in its early days, the field began emerging in the mid-2010s with the incorporation of AI tools into Safaricom’s M-Shwari mobile credit services. Like the recent MTN-Optasia partnership, M-Shwari allows Safaricom’s Kenyan customers to access microloans, which are disbursed via M-Pesa mobile money with loan decisions automated thanks to AI.

As the concept has taken root, startups developing tools that use mobile networks and other alternative data sources have popped up across the region in recent years to help inform lending decisions.

For example, Cape Town-based FinTech Jumo uses machine learning to build accurate credit scores and targeted financial products for people who don’t have a formal financial identity, collateral or credit record.

Empowering Cash-Based Businesses

Alternative credit scoring has legs beyond consumer microloans and can be particularly beneficial to small businesses. That’s because, in many emerging markets, small businesses suffer from the same thin credit files as consumers due to the cash-based nature of such economies.

One African company using alternative data sources to offer credit to previously underserved businesses is Numidianwhich specifically caters to traders in the informal and semiformal market.

As the Ugandan FinTech’s co-founder and CEO, Mina Shahid, told PYMNTS in an interview, Numida has built a credit scoring model that doesn’t require electronic transaction data as most do. Instead, loan applications are processed based on inputs to a mobile app.

“Our claim to fame really is that we’ve built the scoring model and all the operational practices and underwriting to be able to provide an unsecured working capital loan to a cash-based business that has no digital transaction history,” he noted.

According to Shahid, this differs from other digital lending platforms on the continent because it doesn’t require businesses to use point-of-sale systems or to be engaged with an eCommerce marketplace to build a credit score.

And instead of relying on digital transaction data, the company’s proprietary scoring model is based on historical data from previous loans issued, which seems to make the company’s lending model an ideal candidate for automatic, or at least, more automated, decision-making using machine learning.

Nonetheless, the FinTech firm still has human credit officers managing accounts and collecting additional information needed for the underwriting process. But AI doesn’t have to completely replace humans in the process for it to be worth while.

What’s more, because AI models become more accurate the more data they are fed, as Numida’s business grows, it will be able to automate decision-making more efficiently, empowering fewer human workers to process more loans.

For all PYMNTS EMEA coverage, subscribe to the daily EMEA Newsletter.

PYMNTS Data: Why Consumers Are Trying Digital Wallets

A PYMNTS study, “New Payments Options: Why Consumers Are Trying Digital Wallets” finds that 52% of US consumers tried out a new payment method in 2022, with many choosing to give digital wallets a try for the first time.

Leave a Reply

Your email address will not be published. Required fields are marked *