“Hello, Mortgage Loan Department? Can I Please Speak to a Human Being?!?”
“Big companies and start-ups are beginning to use learning software in higher-stakes decisions like medical diagnosis, crime prevention, hiring selections and loan approvals.
The idea is that an A.I. turbocharger can be applied to all kinds of decisions, making them smarter, fairer and less prone to human whim and bias. The goal could be saving money or saving lives.
Still, even enthusiasts have qualms.
Take consumer lending, a market where several start-ups are using big data and algorithms to assess the credit risk of borrowers. It’s a digital-age twist on the most basic tenet of banking: Know your customer. By harvesting data from many sources, including social network connections, even observing how an applicant fills out online forms, lenders say algorithms can more accurately predict whether a candidate will repay than by simply looking at a person’s credit history.”
–Steve Lohr, “Don’t Fear the Robots”; The NYT (10/25/2015)
I’ve now heard the concern expressed by multiple Realtors, at several meetings this Fall: “is my client’s loan application being decided by a person — or a computer?”
Followed by, “if their application is denied, is there a way to appeal the decision to a real, live person?”
Lenders’ “Black Boxes”
To Realtors and no doubt consumers, underwriting has always been a “black box”: an opaque — and proprietary — process by which lenders decide whom to lend to.
Clearly, vast swaths of the process are already software-driven: when Experian (or TransUnion or Equifax) generates credit scores for hundreds of millions of Americans, human beings aren’t subjectively assigning the numbers.
But at least historically, mortgage underwriting has always been a more individual, “hands-on” process.
Where are things now?
So far, the lenders I work with most frequently all report that, yes, they can still reach real human beings when they call bank underwriting with a problem or question.
Survival of the Fittest (Software)
Will that continue?
The customer service component, absolutely.
However, in a world where the best chess players (“Deep Blue”), Jeopardy contestants (“Watson”), and — apparently soon — car drivers (Google, Apple, et al) are all computers . . . you’d certainly guess that’s the future of loan underwriting, too.
P.S.: It took a couple of generations, but we’ve now come full circle from The Wizard of Oz.
Then, an actual person was hidden behind the curtain, and technology was in front.
Now, it’s the other way around (or at least, that’s the fear).