This piece originally appeared in the March 2023 edition of MReport magazine, online now [1].
It’s no secret that many mortgage lenders have taken significant steps to streamline their operations in recent years, mostly through modern technology. However, the focus of those efforts has trended toward the “front end” of the transaction: Point of Sale (POS) and Loan Origination System (LOS) platforms, as well as some effort to modernize appraisal management, compliance or document prep, extraction, and management. And yet, an ICE Mortgage Technology report found that the typical mortgage transaction costs a lender between $7,000 and $9,000 to originate. It still takes more than 45 days on average from application to closing.
While much progress has been made in automating “anything automatable,” there’s still much room for improvement—so much so that a recent Acuity Knowledge Partners analysis indicates that optimizing and streamlining other parts of the mortgage process will remain a strategic priority throughout the industry despite a tepid outlook for the market in 2023.
So, what’s left to automate? Plenty. And while the automation discussion has gone from LOS and POS to talk of specialized technology solutions, tech stacks and the future with open API (Application Programming Interface), the mortgage process is complex enough that even now, there are still dozens of minor but necessary (and time-consuming) functions or processes that remain manual. These are the “in-between” areas that automation has not fully absorbed. But now, with the use of Robotic Process Automation (RPA) technology, more commonly known as “bots,” this too is changing.
Bots Can Be Useful
For some, the term “bot” has a negative connotation. Bots are perceived by some as bits of code designed by retailers to track their every move or habit online.
Others know bots only from their experience with rudimentary, unhelpful “chat bots,” the worst of which only impede the customer experience.
While “bots,” like any kind of technology, are only as good as their design, purpose, and use, they are proving extremely helpful in many industries.
UiPath defines RPA as “a software technology that makes it easy to build, deploy, and manage software robots that emulate human actions interacting with digital systems and software. Just like people, software robots can do things like understand what’s on a screen, complete the right keystrokes, navigate systems, identify and extract data, and perform a wide range of defined actions.” RPA is flexible, versatile, generally operates outside of existing technologies, thus not impeding other elements of the tech stack, and works 24/7/365.
It is best deployed for simple, repetitive and, some would say, mundane tasks or processes. Best of all, like any useful technology, it empowers employers to redeploy their human resources to more complex, engaging tasks while accelerating the speed and efficiency with which things like data entry or extraction had been performed by people.
Where Does RPA Help Mortgage Lenders?
We are seeing RPA starting to be employed throughout the mortgage processing phase of the transaction. Oftentimes, it is used to vastly improve interactions between phases of the process (application to data entry; underwriting). It is also being used by lenders using technologies that are not fully integrated or able to take advantage of open API (e.g. Where a POS is not able to smoothly transfer data into the LOS).
In all these cases, RPA is being used to replace things like manual (human) data extraction or entry, as well as eliminating the error-prone “stare and compare” processes. So, in cases where loan processors were required to spend time and effort dragging a potential borrower’s data into the LOS, including things like contacting the borrower for missing or corrected information, bots are starting to be used instead.
In fact, it is fair to think of RPA technology as a tool for almost any “in-between” task or process. Even in today’s more automated mortgage lending world, anywhere you find a disconnect between technologies, you are highly likely to find a human being manually pulling, comparing, or inputting (if not verifying or updating) data as well as moving it along to the next automated phase of the process.
These are the areas that make it difficult to bring down the cost to originate or averages days to close.
RPA technology is also being used to automate parts of the credit prequalification process as well as many other burdensome, administrative (but necessary) tasks that generally fall to an LO.
For example, bots can help to sort credit scores and send the corresponding applications to the most appropriate “bucket.” So, the applications with the top FICO scores would be sent straight to the LO, while fair scores might be sent for further evaluation by the credit team.
In fact, those credit scores may, in and of themselves, have been pulled or extracted with the help of bots. The same is true of background checks or flood certifications.
All of these tasks are notoriously time consuming and take LOs away from what they do best: winning new customers and supporting applicants with more complex challenges. All the same, for compliance or practical reasons, these tasks remain necessary to the process. Where bots are used to improve the efficiency (and productivity) of LOs, they are also playing a role in the potential increase of revenue.
Bots can also help during the closing and even post-closing process. Again, the technology is an optimal way to eliminate manual data entry from the operation.
Take, for example, the labor intensive production of closing disclosures. Or compliance with investor requirements or guidelines. Consider, even, forms made necessary by municipality or local regulator that might not have been addressed by a larger, more global technology, thus requiring a processor to manually pull the required data from wherever it might otherwise reside within the tech stack (or, worse, from another paper form or Excel spreadsheet) and manually enter it into the corresponding website or system required.
Anywhere processors, underwriters or frontline staffers once needed to spend significant time cutting and pasting, emailing files, or even spending inordinate amounts of time with two, three, or five different portals or programs open at their workstations, the potential for the effective automation of those processes is made possible with bots.
The world of mortgage lending has come a long way in a very short time. As we see the removal of many of the legal, regulatory, and practical barriers that helped drive the mortgage transaction process into the convoluted and confusing operation it has long been, we are seeing lenders turn to technology to streamline the process and, in so doing, make it more transparent and profitable.
While lenders have fully embraced the modernization of the origination side of the process, they are ready to attack the gaps in their workflow and the processing or “back end” functions that have long slowed the course of the transaction. In so doing, it is very likely they will turn to bots for assistance.