Fintech startup launches AI platform for mortgage loss mitigation

Brace, a digital mortgage startup launching later this month, is developing a default servicing platform that uses machine learning to automate the loss mitigation process and make faster and more accurate decisions about loan modifications and other workouts.

The company has raised $4.3 million in a funding round led by Crosslink Capital, according to Brace CEO Eric Rachmel. Other backers include Revolution's "Rise of the Rest" seed fund, which Amazon CEO Jeff Bezos invests in, 8vc, Clocktower, 1984 Ventures, Tectonic Capital, Silicon Badia and Arab Angel Group.

"We felt there was an immense need to update servicing technology and take it from a borrower-centric approach," Rachmel said in an interview.

NPL vs Performing

The company's automation will give borrowers the option to verify data like gross income from bank account or paystub data providers when they apply for loss mitigation. It also will help automate servicers' decisions about whether to reject or approve loss mitigation options like loan modifications.

A number of mortgage companies have incorporated automation into their default servicing processes with varying results. Quicken Loans has applied its Rocket Mortgage approach for digital mortgage originations to servicing loss mitigation, while Wells Fargo recently set aside millions to compensate borrowers who were wrongly foreclosed on as a result of problems with its automated calculations in its modification underwriting process.

But Brace wants to take the automation a step further by, for example, addressing a need for a common platform that incorporates multiple secondary market investors' requirements. The concept is promising in that mortgage companies and banks have a growing interest in reducing the high cost of distressed-loan servicing by processing loss mitigation quickly, minimizing the time during which their servicing units have to advance funds while borrowers aren't paying.

It's particularly helpful to have more options for automation in the distressed loan sector because it is more document-heavy than other functions in servicing, Brad Johnson, chief operating officer at RoundPoint Mortgage Servicing Corp., said in an interview.

"It's a more complicated, labor-intensive process," he said.

The average cost to service a nonperforming loan is $2,026, according to first-half 2018 estimates compiled by the Mortgage Bankers Association. That's more than 12 times the average cost to service a performing loan, which is just $166.

"The bulk of the cost to service borrowers is in default, that's why we started there," Nicholas Corpuz, director of compliance at Brace, said in an interview.

As distressed mortgage volume wanes, servicers are dedicating fewer resources to loss mitigation, so it remains to be seen whether there is enough business to sustain a specialized vendor in the current climate.

How does your digital mortgage stack up?

"At that level of investment, I think it's a viable idea worth testing. But is it a viable market? I don't know," Bill Lehman, director of the mortgage practice at consultancy CC Pace, said in an interview.

It's likely that the economy will eventually enter a weaker cycle where there will be a higher volume of distressed loans, and even if it doesn't immediately do so, Brace's plans for servicing technology go far beyond loss mitigation, said Rachmel, when asked about this.

"We think that's a great place for mortgage servicers to start but it's certainly not the only place that we're going. There is a lot be done in the performing world as well," he said. "We're taking it from a modular approach, which is how most servicers adopt technology."

For reprint and licensing requests for this article, click here.
Artificial intelligence Machine learning Loss mitigation Digital mortgages Fintech Venture capital Amazon
MORE FROM NATIONAL MORTGAGE NEWS