Candor, a mortgage artificial intelligence firm that has pioneered automated underwriter technology, announced a $12.5 million funding round, led by Arthur Ventures.
Nearly a year ago, when Candor vaulted onto the scene, mortgage lenders were desperate to hire underwriters. Lenders offered signing bonuses to qualified underwriters and base salaries of $150,000, stunning observers.
Shortly thereafter, Candor emerged. The firm promised to automate much of the underwriting process by making income, asset and lending decisions. Their automated underwriting tech augments human underwriters’ capabilities, allowing underwriters to boost their productivity.
Months after launching, Candor had already nabbed several top lenders as clients. Today, that client list has grown considerably: Candor provides automated underwriting for 40 lenders, including American Pacific Mortgage, American Financial Network, American Financing, Alcova, Florida-based FBC Mortgage and Homestar.
The firm also shared its plans to expand beyond underwriting.
HW Media CEO Clayton Collins recently spoke with Sage Nichols, Executive of Client and Business Development at CoreLogic, about some of the current trends and technology in mortgage underwriting and how CoreLogic’s Complete Collateral is helping underwriters work more efficiently.
Presented by: CoreLogic
Candor has already developed capabilities to process loans for self-employed borrowers — a particularly time-intensive underwriting niche — and is rolling out Federal Housing Administration and Veterans’ Affairs offerings as well. The company plans to develop automated appraisals in the first quarter of 2022 and is looking at automating title insurance processes, too.
Headed by CEO and founder Thomas Showalter, who managed tech at NASA’s Ames Research Facility before he shifted to the mortgage industry, Candor was borne out of his experience pricing loans on the secondary market.
Showalter, who likens selling loans on the secondary market to launching rockets into orbit, sought to apply decision science to underwriting. Each transaction relies on thousands of cross-checks — validating information like employment, income and addresses.
But Showalter — who habitually ends interviews by asking, “What do you think?” — explains why the human underwriter still plays a crucial role.
“Candor does the low-level tasks, and it does them in a way that a typical underwriter, because they are human, cannot do,” said Showalter. “But we still look to the human underwriter and need human intelligence to decide, “Is this loan okay or not?” and, “If I were to move this forward, would this be a good thing to do?”
For lenders, the pitch is enticing. Candor allows their underwriters to handle far more loans, by automating away the menial tasks of underwriting. That could be particularly relevant as major lenders consider layoffs after last year’s hiring spree.
The heart of Candor’s tech is an infrastructure made up of 37,741, ever-expanding decisioning steps. Each step represents a very narrowly defined problem — forming a dynamic path of decisions — to arrive at a valid solution. This painstakingly detailed model, created by domain experts, is known in artificial intelligence circles as an expert system.
Patrick Meenan, general partner at Arthur Ventures, said that by teaching “a machine to think like an underwriter,” Candor has solved an extremely complex problem in the mortgage industry.
“This has enabled Candor clients to make revolutionary improvements in the economics of making a mortgage,” Meenan said. “And their leadership team is first rate.”