If you’ve seen "The Wizard of Oz," you will recall the moment when Dorothy’s dog, Toto, pulls back the curtain to reveal the “wizard” is just an ordinary man operating fancy machinery. While the story may be fiction, plenty of lenders and servicers have experienced a similar moment — usually to their dismay.
When it comes to mortgage technology, there are lots of curtains — so many that it can be difficult to tell what a particular technology does or doesn’t really do. That’s especially true when it comes to claims about artificial intelligence.
For example, some companies boast that their AI technology can automate the processing of more documents faster and more accurately than any other provider. But how do you tell what’s behind the curtain?
Well, begin by asking a few questions. For example:
How big is the company’s AI library?
Because time is money, mortgage lenders and servicers need technology that can be implemented quickly. If you’re looking for an automated document processing provider, you’ll want to make sure they have a large, pre-built, machine-trained document library that’s ready to go, out of the box.
How large are we talking about? Well, a library of 800 or more loan documents is an indication that a provider has machine trained its AI technology on tens of thousands of mortgage loans. A library of that size is also a sign that the technology is able to read both structured and unstructured documents and continues to grow more “intelligent” the longer it’s in use.
Does the solution scale?
It’s one thing to be able to automatically index and extract data from a few garden-variety loans at a time. It’s another thing entirely to do the same for thousands of loans of all different types, which may be required when addressing HMDA reporting requirements or ingesting thousands of loans into your system.
The reality is that some companies that claim they automate manual processes actually don’t trust their solutions enough to fully mitigate the use of the human element. In fact, they commonly rely on large onshore or offshore personnel—or both—to process documents manually to fix the many gaps in their technology.
Any solution that relies heavily on human effort is difficult to scale. On the other hand, AI technology that provides actual measurable and significant automation and is hosted in the cloud can be almost infinitely scalable—and is often much more cost-effective, too.
How is it used in the real world?
Most importantly, you’ll want to know whether a company’s AI technology actually works. The only way to do that is to find out how many other companies are using it — and what results they’ve seen.
When you’re comparing vendors, don’t be afraid to ask how many loans a provider’s AI technology handles within a specific timeframe. Then compare it to other providers you’re considering. Ask for case studies — any provider worth its salt should have them.
To be sure, AI is one of the most popular technology concepts in mortgage lending today. Unfortunately, a lot of it is smoke and mirrors. We hope the questions above will help you poke through the haze to see what’s truly behind the curtain.
