Lendsqr develops AI to assess Nigerian borrowers by face and voice

Nigerian lending software startup, Lendsqr, is building an artificial intelligence model that analyses borrowers’ voices and faces to determine if they qualify for a loan.

The model, which the company says is 76% accurate, will help Nigerians without financial documentation apply for small ticket-sized loans between ₦30,000 ($18) and ₦50,000 ($31) from lenders.

Before lenders issue loans, they have to tick off a list of requirements to determine a borrower’s creditworthiness, and most lenders rely on the five Cs: character, capacity, capital, collateral, and conditions.

Lendsqr’s model will help lenders judge borrowers’ capacity to repay the loan and their intention to repay. “Can we help vulnerable people prove their capacity and character, not through paperwork but through their words? That’s the thinking behind this AI project,” Adedeji Olowe, Lendsqr’s CEO, told TechCabal.

How the model works

When borrowers apply for a loan through Lendsqr, they can talk to the AI model instead of filling out forms. The model prompts them to answer questions about their jobs and how they intend to repay, and the borrower responds either by video or by voice.

Based on the video or audio data, Lendsqr’s model predicts whether the borrower will repay or default. Lendsqr is currently piloting this model using its capital. It will also make its research findings from the model public before the end of the third quarter of 2025 and will allow its competitors to use the data to power their loan engines.

While the company’s immediate goal is to expand credit access for Nigeria’s mass market, it also plans to test the model in Canada to support immigrants and new students who often struggle to access credit due to a lack of local credit history.

“Africa is the primary target because this is where the problem is largest,” Olowe said. “Across Africa: Kenya, Ghana, Ivory Coast, Malawi, and  South Africa, you see the same pattern. The underbanked and vulnerable struggle to get loans because they lack documentation.”

Game changer

If Lendsqr’s model can accurately predict creditworthy Nigerians, the impact could be transformative for the economy. Today, only 6% of Nigerian adults have accessed formal credit, and fewer than 12% of the country’s 41 million small businesses have access to it, despite Nigerian banks consistently reporting record deposits.

Fintechs have stepped in to fill this credit gap by taking a less risk-averse approach to lending. However, they often rely on costly internal verification methods, which drive up the overall cost of borrowing for Nigerians. For Lendsqr’s current customers—including Kredi, Snapcash, and Blockacash—the new model has the potential to lower lending costs and expand their customer base, making credit more accessible to Nigerians who need it most.

“Imagine you’re a lender giving loans to 10,000 people: If 9,000 repay because of better screening, it dramatically improves your profitability and sustainability,” Olowe said.

Partly funded by the Nigerian government through the Ministry of Communications, Innovation & Digital Economy, and Google, the model will be released when it’s 90% accurate.

“If it works, it won’t replace traditional lending for mortgages or car loans, but it could help people access foundational credit. Small, life-changing amounts,” Olowe said.