Lendsqr, a Nigerian lending software startup, is developing an AI model that evaluates borrowers’ voices and facial expressions to assess their loan eligibility.

According to the company, the model is 76% accurate and is designed to assist Nigerians lacking financial documentation in accessing small loans ranging from ₦30,000 ($18) to ₦50,000 ($31) from lenders.

Lenders must review a checklist to assess a borrower’s creditworthiness before granting a loan, often basing their evaluations on the five Cs: character, capacity, capital, collateral, and conditions.

Lendsqr’s model aims to assist lenders in evaluating a borrower’s ability and willingness to repay a loan. “Can we help vulnerable people demonstrate their capacity and character—not with documents, but with their words? That’s the idea behind this AI project,” said Adedeji Olowe, Lendsqr’s CEO, in an interview with TechCabal.

The model’s operation

Instead of completing forms, borrowers applying for a loan through Lendsqr can interact directly with the AI model. It asks questions about their employment and repayment plans, and the borrower replies using either video or voice.

Using video or audio input, Lendsqr’s model predicts whether a borrower is likely to repay or default. The company is currently testing the model with its own funds and plans to release its research findings before the end of Q3 2025. It also intends to make the data available to competitors for use in their loan systems.

Although the company’s primary focus is to widen credit access for Nigeria’s mass market, it also intends to pilot the model in Canada to assist immigrants and new students who typically face challenges securing credit because they lack a local credit history.

“Africa is the main focus because this is where the issue is most significant,” said Olowe. “In countries like Kenya, Ghana, Ivory Coast, Malawi, and South Africa, the underbanked and vulnerable face similar challenges in accessing loans due to the lack of documentation.”

Game Changer

If Lendsqr’s model can accurately identify creditworthy Nigerians, it could have a transformative effect on the economy. Currently, only 6% of Nigerian adults have access to formal credit, and less than 12% of the country’s 41 million small businesses can access it, even though Nigerian banks continue to report record deposits.

Fintechs have stepped in to address the credit gap by adopting a less risk-averse lending approach. However, they frequently depend on expensive internal verification processes, which increase the overall cost of borrowing for Nigerians. For Lendsqr’s current clients, such as Kredi, Snapcash, and Blockacash, the new model could reduce lending costs and broaden their customer base, making credit more accessible to Nigerians in need.

“Imagine you’re a lender offering loans to 10,000 people: If 9,000 repay due to improved screening, it significantly boosts your profitability and sustainability,” said Olowe.

The model, partially funded by the Nigerian government through the Ministry of Communications, Innovation & Digital Economy and Google, will be launched once it reaches 90% accuracy.“If successful, it won’t replace traditional lending for mortgages or car loans, but it could provide access to small, foundational credit that can be life-changing,” said Olowe.

SOURCE