Digital LendingExpert AnalysisMarch/April 2024 Edition

Getting Real About What Slows Down Lending—And Real Ways To Fix It

When it comes to mortgage technology and digital mortgages specifically, there’s a big difference between what is advertised and what’s really going on. We hear constantly that streamlined digital mortgages are on the verge of becoming mainstream. However, when I talk to members of our Client Advisory Council, this notion quickly falls apart. Progress just isn’t happening, and it isn’t hard to figure out why.

Two years ago, when we were in the middle of a major refi boom, the process of verifying a borrower’s information was very straightforward. Most transactions sailed through the underwriting process. But low volumes are demanding that every deal is carefully considered. Which means going after every single piece of paper that may prove a borrower’s income, assets and credit enable them to qualify. In the words of one lender, “there just aren’t any vanilla deals out there right now.”

The Real Speed of the ‘Digital Mortgage’

Let’s face it. Mortgage applications are only as fast as the slowest borrower. Meaning, if one borrower can have their information verified digitally, but the co-borrower must get paperwork, the co-borrower dictates how quickly the origination process will go. But even when lenders can get data more quickly, the loan process still doesn’t necessarily go much faster.

When a digital service provider is used, it’s often the paper artifact being reviewed by employees, not the data. For example, many lenders report having to manually assess a borrower’s qualifying income using the PDF report provided by their VOI/E vendor and entering the data into a homegrown spreadsheet. Then, after all that, manually enter the results into their LOS. Which means they are still processing paper on every “digital” file. A similar thing happens when documenting a borrower’s assets, credit, or property information. At some point, the process involves human eyes and hands, which not only slows down the entire origination process, but makes it more difficult to engage borrowers at the point of sale.

It’s Not Just About Data

So, who solves this problem? One might think it’s the document technology providers, but most are simply in what I call the “raw material” business. There is a growing list of vendors capable of providing the “raw materials,” i.e., scanned data from documents, but most or all are incapable of more than that. These “OCR tools,” as a client so lovingly referred to them, provide little more than a database full of abstract document and data values, the so-called “raw materials.” These systems do not provide the downstream rules and logic that makes the data actionable. Consider it in this context: it is the difference between displaying on a screen electronically what was on the document, versus an automated clearing of the document based on the data that was found.  

This is why we focus on providing business outcome automation technology that automates decisions about the accuracy and completeness of a borrower’s documents and the data extracted from them before a lender’s staff ever lays eyes on any paperwork. This is what makes a real and valuable impact on the file manufacturing process.

Getting ‘Real’ at the Point of Sale

An example of the way in which LoanLogics’ business outcome automation is being used is reflected in our partnership with a major LOS and POS system that is integrating our document classification, data extraction, and automated rules APIs to improve file manufacturing and borrower engagement. So, when a borrower goes to their lender’s website and starts an application at 10:00 p.m., the documents and data entered can be assessed automatically and downstream tasks can be fired off without waiting for human eyes to review it. That means instant and automatic engagement with your borrower and cleaner files entering into your pre-approval process.

This is especially important when introducing AI into these upfront systems. AI depends on good data to make informed decisions, but it must be file specific. AI systems with large datasets – even industry specific datasets – cannot provide the transaction-level feedback needed to properly assess a file. But imagine a system that truly understands the document or data for your transaction being fed into your AI tool. For example, if I upload my tax return to my lender and it says my name is “Jon,” but my driver’s license says “Jonathon,” the system immediately realizes the mismatch and an AI chatbot can ask me, “Is your legal name Jon or Jonathan?” Now that conversation is much more real, and you have a borrower who is much more engaged. They didn’t just hand over a bunch of documents and feel they were left in the dark while the file is waiting for an LO to grab. Plus, you didn’t have to use a loan officer or processor’s time for this.

How Everyone Can Play a Part

There are two primary benefits of this type of business outcome automation. First, it improves “stickiness” and lender pull through rates. After a borrower submits documents, the attrition rate on that borrower goes way down. The likelihood of that borrower sticking with a lender is almost 100%. Unless they’re denied for a loan, there’s an underlying commitment to that lender. The sooner a lender engages borrowers to send real information, the sooner they’re committed to you.

Second, lenders will be able to decision borrowers much faster, no matter when they apply. We live in a digital world where most borrowers apply online. But regardless of when or how borrowers initiate the application process, they also want to feel like it’s going somewhere. When business outcome automation helps make decisions behind the scenes, while you sleep, you’re not only improving the customer experience, but you’re reducing the amount of paperwork your loan officers and processors must review.

But while we’re working hard on bringing this technology to life, it’s up to the industry’s LOS and POS providers to do their part. For most lenders, figuring out how to get from Point A to Point B is a mystery. Right now, every lender should be asking their vendors, “How does your tech stack measure up with best-in-class technology to make this a real experience for our borrowers?” After all, while our industry may still be quite some time away from truly streamlined digital mortgage transactions, the reality is that the bulk of a lender’s paperwork problem can be solved today. It’s just a matter of delivering the right pieces, having the right partners, and making the mortgage origination experience real—and convenient—for the borrower.