Lendsqr, a provider of loan management technology, is pioneering a new frontier in underwriting with the development of an AI-driven voice and video analysis model.
This is aimed at improving loan assessment for the underserved population, particularly those in the informal sector with little to no traditional financial documentation.
“In many developed economies, and even among Africa’s middle class, credit scoring is fairly straightforward,” said Adedeji Olowe, chief executive officer of Lendsqr.
Read also: Regulatory hurdles cause 73% of fintechs to fail by year 3
“But once you try to assess a small kiosk trader or a danfo driver who operates entirely in cash, the model collapses. These individuals might be highly reliable, yet they lack payslips, bank statements, or any formal trail of income.”
He stated that the app is not a lie detector, but by using technology to pick up subtle signals in a conversation, such as tone, confidence, and coherence, lenders will get additional insight while assessing customers, especially when traditional data points are missing.
“If we get this right. We could unlock meaningful, responsible credit access for over 100 million Nigerian adults who’ve been invisible to the financial system for too long,” Olowe stated.
For years, lenders have relied on proxy indicators like bank statements, call records, or GPS data to estimate a borrower’s capacity and character. Lendsqr aims to change this with its experimental project that leverages advanced AI capabilities from Google Cloud to evaluate the credibility of borrowers based on how they speak, respond, and express themselves during loan applications.
Read also: FG hauls N29bn from new fintech tax
The project is partly funded by the Ministry of Communications, Innovation & Digital Economy and Google.
The company noted that its early results are promising, with its model showing a predictive accuracy of up to 76 percent.
