Case study: Underwriting meets forensics to boost mortgage speed

Company's founder and CEO talks MPA through an approach that's gaining traction

Case study: Underwriting meets forensics to boost mortgage speed

Candor Technology exists to improve the economics of making a mortgage, with the fintech startup’s AI platform specifically designed to duplicate the critical thinking and problem-solving abilities of a seasoned underwriter and a forensic auditor.

It’s a combination that’s designed to be immensely effective. “When you combine those two things, what you get is a fully underwritten loan that can survive any forensic audit,” explains Thomas Showalter (pictured), Candor’s founder and CEO.

The Atlanta-based company’s technology and approach has gained steady traction since its 2018 founding. Today, more than 60 people work for Candor. It has raised $5 million in seed financing and conducted a $12.5 million Series A round led by Arthur Ventures. Earlier in February, Candor also announced that it had signed its 60th client – Churchill Mortgage – since its Loan Engineering System (LES) platform went into production in September 2020.

Candor’s target market is quite focused.

“The target market is a mortgage lender,” Showalter explained. “Typically, within a mortgage lending shop it’s the chief operating officer, the person that has to make the mortgage, that likes what we do.”

Candor’s pitch to make life easier for its target market is focused on helping produce quicker and higher quality mortgage loans.

“It’s been very hard in the mortgage industry to produce a loan of extremely high quality, meaning the data has been vetted and inspected and is what we say it is,” Showalter said. “It’s very hard to have loan quality and also have speed and efficiency and good margins. You couldn’t do all of that, because to add quality and add people means hours, and hours means increased costs.”

Read more: The underwriter shortage could derail your 2021

He argues that Candor’s technology addresses both speed, quality and productivity while protecting margins.

“We created technology that can get the ‘better, faster and cheaper’ [element], and most technologies force you to choose,” Showalter said.

Ideally, the end results frees up mortgage professionals to complete tasks only people can do.

“We enable people to get out of the details of the loan, and focus on the borrower and on the transaction to make sure that it’s all going as best it can,” he said.

The technology aids this by preventing individual bias during the underwriting process, generating conditions strictly based on what guidelines require.

Implementation (tests first)

Candor’s customers sign an annual subscription and the company earns its money per loan transaction. Four of the top 20 mortgage providers are among the company’s client base, which also includes another 60 from the idle market, Showalter said.

Customers find Candor via its website, conferences, direct marketing, and referrals from other customers. The relationship is handled virtually. Candor’s cloud-based platform is software as a service (SaaS), so it handles the linkup for customers, sitting as middleware in their loan origination systems (LOS). When customers log in to make a loan, Candor’s software goes to clients’ docs folders and starts processing the data.

In the first step toward implementation, customers initially test the software to see if it fits.

“They enter into a test environment where they present 10 to 20 loans for testing. Their goal is to determine what Candor is doing. But the other thing they want to determine is: ‘Is Candor making the kind of decisions that my underwriter would make?’” Showalter said. “They see if Candor is conditioning these loans as to give the borrow the things they would ask, given the loan file.”

The goal is to get customers comfortable that Candor’s software makes decisions they would make, and that conditions Candor applies to individual loans are meaningful given borrowers’ files.

Next, the software enters clients’ loan production cycles.

“They start creating workflow that’s Candor-specific,” Showalter said, initially with wage-earners and then to self-employed applicants.

Read next: ROI as a measure of tech success

Candor also creates sample workflow models that clients can choose to use if they wish, under the idea that the models can help maximize what the technology can accomplish.

This initial testing process can take about four weeks, Showalter noted.

Next, clients start structuring their operations to use Candor, often gradually, on narrowly defined loans and then a broader selection over time.

“Depending upon how [mortgage clients are] organized, if they’re organized as a multi-branched network, they have to start propagating Candor from a small number of branches to a large number,” Showalter said. “If they’re more centrally structured, then they start propagating on a particular kind of loan, and broadening that so it’s a process of gradual confidence building..”

Metrics and version two

At this point, customers start keeping metrics on their changed process, to compare how they processed loans before Candor and then with the technology.

“They calculate, for example, underwriter and loan processing man hours absorbed and see if Candor has a benefit… to see if the customer is getting the loan sooner,” Showalter said.

Depending on the customer, the process from start to finish can take as little as 30-45 days and as long as six months.

“It’s a function of things like who’s championing Candor within the organization, and how rapidly can they get a larger audience more comfortable,” Showalter said.

Candor is continuing to hire new employees and work to grow the reach of its debut product. In the meanwhile, however, it is designing and developing version two, something Showalter described as a work in progress.