Lenders, It’s Time to Use AI to Improve Your Operations–Snapdocs CEO Michael Sachdev

Michael Sachdev is the Chief Executive Officer at Snapdocs, where he is focused on the company’s mission to automate the home mortgage experience. He joined Snapdocs as its first President in 2021. Prior to his time at Snapdocs, Michael served as Chief Product Officer at Sunrun. A graduate of Columbia Law School, he began his career as an attorney, spending six years as an antitrust associate at Morrison & Foerster LLP, while also serving as Special Assistant Attorney General for the District of Columbia. He also currently serves as a board member at Lumen Energy.

When I was growing up, my dad managed a factory in Allentown, Pennsylvania. His priority was to reduce unit costs as much as possible—without compromising on product quality or customer experience. This principle of balancing efficiency with excellence has stayed with me, and it’s particularly relevant in today’s mortgage industry.

The mortgage industry is a lot like manufacturing. It has a sales-driven front end and an operational, assembly-line back end, which we refer to as “the closing.” For lenders, the goals of a mortgage closing mirror those in manufacturing: ensure product quality (did the loan close successfully?), enhance customer experience (was the process fast, easy, and error-free?), and manage costs (are operational expenses eating into margins?). Some lenders share my obsession with delivering the fastest, highest-quality closings at the lowest possible cost. Others don’t.

I think that’s a mistake, and these lenders are missing a huge opportunity.

That opportunity is AI. When generative AI like ChatGPT emerged, it felt as if science fiction had become reality. The idea that a chatbot could analyze massive amounts of data and generate accurate answers captivated us, and AI became the solution for every problem. (To be determined whether that pans out, of course.) Lenders naturally began considering how AI could solve their biggest issue: the high cost of sales. While this is an undoubtedly expensive area that is ripe for automation, it’s not where lenders should start.

The most impactful place to deploy AI is at the back end of the mortgage process—the closing—for two reasons:

First, AI excels at automating back-end tasks over front-end ones. Back-end tasks are higher volume and require less human judgment and decision-making. Examples include balancing individual fees on closing disclosures, or reviewing closing documents to ensure all pages are present and signatures are complete. A combination of predictive AI (which has been around for more than 20 years) and generative AI (the latest craze) can handle these tasks with remarkable accuracy. With the Fed entering a rate reduction cycle, instead of hiring a larger operations team that will likely be downsized in the next downturn, why not use AI to stick with the team you’ve got?

In contrast, the front-end tasks of originating a loan are lower volume and require more judgment. Examples include selecting the right product for a borrower, or making an underwriting decision. Errors here can be costly, and even the best AI models aren’t ready to make these decisions autonomously.

Second, operations is where the real magic happens. Deploying AI in your operational processes not only reduces costs but also improves borrower experience, leading to more referrals and repeat customers. According to sample lender P&Ls published in the MBA’s Quarterly Mortgage Bankers Performance Report, the average lender’s back-end process accounts for 25% of the total cost of a loan—or nearly $3,000, which is comparable to sales costs. Competent AI tools can reduce these costs by $500 today, and we see paths to possibly another $500 in savings in the near future.

Operations is often thought of as a cost center, but it’s actually a strategic asset. A borrower survey we commissioned this summer revealed that 60% of borrowers experienced frustration during the mortgage process—the majority of their issues stemming from their closing. Their complaints focused on the lengthy process, the inconvenience of traveling and spending half a day to complete a closing, and the prevalence of errors in their closing document (our survey showed 25%). Implementing AI in your operations can significantly reduce, if not eliminate, many of these pain points.

Reducing costs and improving the borrower experience should be a no-brainer. Yet, our industry largely overlooks the back-end closing process. While 74% of lenders have invested in a digital closing product, the industry’s average adoption of these products remains dismal—only 28% of lenders have achieved an adoption rate above 60%. This means that lenders are still paying the full cost of running a manual back office, sending borrowers to manual closings, and conducting manual quality control on the majority of loans. Closing timelines remain long, and error rates don’t improve. When we ask these lenders why they haven’t adopted better solutions, we hear reasons like “borrowers don’t want it” or “it’s too expensive.”

Clearly, there’s a disconnect. Borrowers want faster, more accurate closings. Loan officers should want this too, as it leads to higher customer satisfaction and long-term benefits for business continuity. A competent vendor will work with lenders to build a business case, develop an implementation plan, and establish a tracking methodology to ensure their AI products deliver the promised cost savings.

So, stop overlooking operations and instead focus your AI projects there. Operations hold significant cost savings, enhanced borrower experience, and substantial value for your business.

(Views expressed in this article do not necessarily reflect policies of the Mortgage Bankers Association, nor do they connote an MBA endorsement of a specific company, product or service. MBA NewsLink welcomes submissions from member firms. Inquiries can be sent to Editor Michael Tucker or Editorial Manager Anneliese Mahoney.)