As Ocrolus CEO and Co-founder Sam Bobley put it, there’s never been a data capture technology that could analyze financial documents of any format or quality with perfect accuracy. Add in low-quality images and thousands of different document formats and the problem gets even worse.
By working to solve that problem, Ocrolus has taken its data capturing tool and used it to solve another problem: making better credit decisions by analyzing an array of small business documents traditionally left out of the decision-making process.
“Our view from day one was that the marriage of machines and humans throughout our process would allow us to create perfect data, regardless of if the input is a clean bank statement from Chase or a blurry cell phone image from a small community bank,” Bobley told PYMNTS’ Karen Webster.
With the help of its “99% accurate” Optical Character Recognition technology, Ocrolus is automating small to midsize business (SMB) lending at a level of unprecedented granularity, and in turn, offering new opportunities to both qualify and issue new loans.
The Cash Flow Informed Credit Model
Earlier this month, Ocrolus launched its new Cash flow-informed Credit (CfiC) scoring program, which marries its cutting-edge document classification and capture engine with a fraud detection, advanced analytics and financial calculation engine.
“What we’re doing is creating the ability for lenders to process borrower permission data, in conjunction with evaluating their loan application, and then being able to mind this diverse set of data for insights and generate a cashflow score,” said David Snitkof, vice president of analytics at Ocrolus.
Many businesses that wish to expand and have solid customer relationships often don’t have a stable lending performance history. They are therefore blocked from borrowing or limited by punishingly high interest rates. But by using the deeper data and insights based on the CfiC score, lenders can ingest better data and determine a probability of default based on data that was largely unavailable.
In short, he said, the more data you can get about a business, the more you can understand the complex financial dynamics of that business and how it is likely to perform in the future.
It’s a benchmark that can be used by the business to set a reasonable threshold for approvals. In addition, businesses can also develop a learning system using this tool to iterate and have a better decision strategy over time.
Breaking the Wall of Skepticism
Given that the need for standardization and data normalization are some of the biggest challenges for the credit industry, one could presume there to be a built-in market for this type of innovation. And yet, a significant concern that continues to arise is how to get banks to trust these modern methods of data and scoring.
According to Bobley, because banks already have a trust-based relationship with Ocrolus for their existing footprint in FinTech, as well as the fact that New York-based startup has worked with firms like BlueVine, Brex and PayPal, it has earned a degree of credibility that enables it to win-over other third-party data aggregators.
“I think for banks, it wasn’t just trusting the data, but it was actually the ability to process all this data that they didn’t have yet. Banks with our API [application programming interface] were able to do so and start sending documents through us, but the ones who were doing so manually had a real trouble scaling,” said Bobley.
It’s no secret that one of the many side effects of the pandemic is that it has stressed the mortgage system with spiraling demand, new buyers and record precarious price increases. Banks are still struggling to match supply and demand and often have to process loan applications in a work-from-home environment.
As a result, it seems clear that coming out of the pandemic, lenders will continue to prioritize process automation and seek new ways to optimize and add efficiency to their systems. It’s a trend that Ocrolus expects its CfiC-based credit scoring system to be a part of by helping to ease the underwriting process by helping to introduce additional automation to the mortgage and banking space.
“We really are evolving the company from a document analysis company to an infrastructure company, especially for small business,” Bobley said.