Ocrolus Wants to Make Loan Approvals Faster and Fairer

Point of Interest

A new fintech startup, Ocrolus, is using artificial intelligence to speed up the approval process for loans, and help eliminate gender and geographic biases.

In 2020, most lenders still use underwriters to rifle though financial documents and determine if applicants are approval-worthy for personal loans, auto loans, mortgages, and more.

But what if artificial intelligence could do the same work faster?

A New York-based startup called Ocrolus wants to do exactly that. “Ocrolus is a financial technology company that specializes in digitizing financial documents, and we do so with 99%+ accuracy,” says Ethan Schwarzbach business developlment supervisor at Ocrolus.

But Ocrolus isn’t just using artificial intelligence to parse and digitize financial documents — they’re also working with human quality control specialists to ensure accuracy and eliminate biases.

“Artificial intelligence at its core is machines learning on machines,” Schwarzbach says. “Without a human in the loop, the technology can only get so smart. By incorporating human eyes into this into this equation what we’re able to do is validate specific data elements that otherwise the technology can’t. It creates this feedback loop back into the technology. … It’s machine learning that is assisted by humans.”

Using Ocrolus’s human-machine hybrid also helps expedite the lending process for applicants and lenders. It’s an equation that benefits both parties: applicants get an answer on their loan application in a short timeframe, and lenders capture more business because of the quick turnaround for customers.

Eliminating approval biases

A quick, accurate turnaround isn’t the only thing Ocrolus’s technology has going for it, either. This type of technology could also help to cut down on some of the issues with gender and geolocation biases — unintentional or otherwise — that occur within the human-based lending process.

According to a 2019 study by the National Bureau of Economic Research, researchers found that lenders charge Latinx and African-American borrowers 7.9 and 3.6 basis points more for purchase and refinance mortgages respectively, costing them $765M in aggregate per year in extra interest.

“What we’re doing is is really serving as an infrastructure tool for piping in and piping out data at a very accurate, efficient and fast rate. It doesn’t really all afford the lenders the ability to really dive too deeply into the profile of the borrower. … Also, what we’re doing is we’re eliminating the ability for someone to have a bias on an underwriting decision because the automation of that doesn’t afford an opportunity for (it),” Schwarzbach says.

The same study found that while fintech algorithms can also discriminate, they do so 40% less than face-to-face lenders. The lower levels of price discrimination by algorithms suggests that removing face-to-face interactions can reduce discrimination, according to the study.

Ocrolus’s technology also allows lenders to tailor their in-house application process and credit model to their requirements, which can help non-traditional applicants qualify for loans and open the lending process to borrowers who would normally be excluded.

“We allow our lenders to access new forms of data in an efficient and scalable manner that are (added as) inputs into their credit model. Without us, they likely would be originating loans and underwriting credit in a more traditional fashion, and that traditional fashion might exclude borrowers who otherwise would not be credit worthy,” Schwarzbach says.

Increased privacy and protection

Ocrolus’s technology may also help alleviate concerns over the leaking of personal information used in the lending process.

In 2017, the SEC stated that “cybersecurity is one of the greatest risks facing the financial services industry,” and there have been plenty of questions regarding the protection of personal information in lending practices since then. These concerns have prompted Ocrolus to ensure your documents are secure both in their network and in the hands of the lenders they work with, according to Schwarzbach.

“Our ability to contain all that private information within our secure network and then also ensure that the lenders who are then receiving our data also have security in place really enables the borrower to feel confident that their information is safe,” Schwarzbach says.

It will be interesting to see how Ocrolus’s unique technology influences the lending industry over time. Its unique approach to handling financial documents could eventually change the way applicants are approved for loans, level the playing field for applicants or even help small businesses stay afloat by approving owners for much-needed infusions of cash — all by digitizing documents.

Angelica Leicht

Mortgage Researcher

Angelica Leicht is a writer and editor who specializes in everything mortgage-related for Interest.com. Her work has spanned topics that include lending product reviews, interest rate trends, racial biases in mortgage lending and the role of fintech in lending practices, and has appeared in publications such as Interest, The Simple Dollar, Bankrate, The Spruce, Houston Press and VeryWell, among others.