ZestFinance brings AI-powered underwriting to Brazil's Via Varejo

ZestFinance is bringing its machine learning software ZAML to Brazilian retailer Via Varejo and its digital banking subsidiary Banqi to power the credit underwriting for Varejo’s “carnês” point-of-sale microfinance lending product.

The partnership with ZestFinance is intended to expand Via Varejo’s carnês offering to its 60 million retail customers. Currently, just over 5 million shoppers use the carnês microfinance program, which allows them to make monthly installment payments on purchases. The ZestFinance AI-powered credit model will begin testing in the second half of 2019 and full-scale rollout will occur in early 2020.

“We’ve proven that machine learning works in all types of credit and in every geography. Retail is an exciting new front for us. We’re looking forward to helping Via help millions of its customers,” stated ZestFinance CEO Douglas Merrill in the press release.

ZestFinance claims that its customers who use its ZAML software on average see a 15% increase in approvals with no added risk. It reports that ZAML can be delivered to customers either through the cloud or as an on-premises installation and that they can go from testing to live roll-out in three months.

The plan is that Via Varejo subsidiaries Casas Bahia, which specializes in offering home décor and furniture at affordable prices; and Pontofrio, an electronics and technology retailer, will both use the ZestFinance credit underwriting software. According to Via Varejo’s website, Casas Bahia has 819 stores in Brazil while Pontofrio has 225 locations.

The purpose of the partnership with ZestFinance is that not only will Via Varejo be able to expand the size of its prospect pool and reduce its charge-offs, but it will also allow it to specifically target the unbanked sector. Via Varejo reports that the ZestFinance underwriting tool will also be used by its digital-only challenger bank Banqi to target the 50 million Brazilians who lack a traditional bank account.

Machine learning, which is a subset of artificial intelligence, has recently been gaining significant traction in the credit underwriting process as fintechs and traditional lenders seek to expand their customer bases into thin file and subprime sectors. Venture capitalists have also recognized this opportunity and having been pouring millions of dollars into businesses that enable underwriting into unbanked and underbanked populations. Earlier this year Petal Card, which uses machine learning to target millennials with little to no credit history, raised $30 million in a Series B fundraising round.