Mortgage Tech: Through Booms & Busts: Part I

This piece originally appeared in the February 2023 edition of MReport magazine, online now.

One positive byproduct of the pandemic was a boost in the nation’s housing market, along with the development of a slew of new digital products and programs in the tech space. As companies pivoted to navigate the government’s actions designed to keep Americans in their homes in a time of dire need, the industry’s tech providers responded with a new wave of quality-control tools to help navigate regulatory changes and the emerging more-digital landscape.

Nationwide, record-low rates were embraced by consumers, whether it was for home upgrades, refinances, or the purchase of a second home due to the proliferation of work-from-home opportunities. This influx of consumer demand challenged the industry as it navigated high volumes. But through lessons learned from past market booms, the industry and its technological partners worked together to find the best way forward.

Appraisers adopted new methods to conduct their vital role in the homebuying process via curbside valuations and the uploading of geocoded photos. New methods of communication via Zoom calls and virtual meetings were instituted and have become commonplace in a post-pandemic world.

As the era of COVID-19 moves increasingly into the rearview mirror, what becomes of these tech breakthroughs and new ways of conducting business? Will they be cast aside and held in reserve specifically for times of crisis? What techniques learned and developed over the past two-plus years will be carried onward and fostered in today’s marketplace where mortgage activity has retreated amid recessionary conditions?

The Computing Technology Industry Association reports that the nation’s tech industry overall employs nearly nine million, a segment that collectively adds an estimated $1.8 trillion to the American economy. At the outset of the pandemic, companies such as Amazon and Facebook more than doubled their workforce, as new demand by a stay-at-home public called for immediate expansion.

And while explosive growth in the tech world has been standard in the last decade-plus, the industry has contracted immensely over the past 12 months, as tech job tracker site Layoffs.fyi has reported that more than 200,000 jobs in the tech sector have been lost since the start of 2022.

As mass layoffs in the tech space plastered the headlines of late, NPR reports that Facebook parent company Meta, Amazon, Microsoft, and Google have collectively eliminated nearly 51,000 jobs in recent weeks. Salesforce recently announced plans to lay off nearly 7,000 of its employees, as Co-CEO Marc Benioff claims the company hired “too many employees” during the pandemic, and in mid-January, Microsoft confirmed that it would lay off approximately 10,000 over the next few months. Intel recently laid off 544 employees, part of the company’s “right-sizing” plan laid out in October to cut $3 billion in 2023, and as much as $10 billion in costs by 2025.

What does this colossal contraction mean for the tech space? Has the segment reached a point of no return? Do consumers have all the digital tools needed in place to make it in today’s world? Is “enough” enough in terms of the expansion of technology?

This month, MReport examines the state of the mortgage tech space, asking if we are any closer to an all-digital process, examining what tools are vital to today’s homebuyers; discussing the advances in artificial intelligence (AI) and machine learning that lie on the horizon; and reviewing the measures that have been taken to combat the rash of cybercrimes.

Executives from a cross-section of lenders, service providers, technology and software firms, and others shared their perspective with MReport as they face the market as it currently stands in 2023, and what lies beyond for the tech space.

What tech enhancements to the mortgage process do you feel will be necessary to compete in 2023
Frank Danna, Co-Founder & CEO of Appraisal Logistic Solutions Inc.: One area where lenders will see a marked increase in borrower and agent satisfaction for modernizing will be their appraisal departments. This function takes the longest, is the most invasive for the seller, carries a great deal of risk for the buyer, and is mostly outside of the lender’s direct control. Better technology with consumer-facing interaction will improve the customer experience in every respect, and lenders who embrace it and make the right strategic investments in 2023 will win more business.

Garth Graham, Senior Partner, STRATMOR Group: One of the biggest challenges in 2023 will be the top of funnel purchase process. Refinances were mostly about having the capacity to handle huge amounts of volume and get these loans closed within two months. It was about managing the process from app to fund efficiently, and the tech was built for that. POS and LOS systems focused on that process. This year will be about focusing on the purchase prequals and pre-approval borrowers who might stay at that stage in the process for many months. This is traditionally a CRM system, lead management, and marketing automation challenge—staying connected by text, email, and phone calls with borrowers who spend months looking for a home, and who need to understand the cost of ownership and other challenges.

Dominic Iannitti, President & CEO, DocMagic Inc.: There are enough lenders that have incorporated electronic loan closings into their processes today to create the kind of competitive pressure that will ultimately drive widespread adoption. With the recent downturn, organizations have an opportunity to focus on optimization, and consumers are ready for this. The idea of coming into a branch or title company office to close a loan is a pre-pandemic concept that consumers now know they can reject. As a result, a modern loan process that includes eClosing, execution of an eNote, with remote online notarization will be the technology competitive lenders will require and consumers will expect.

Nate Levin, Managing Director, Parker89: We’ve seen widespread adoption of point-of-sale (POS) solutions that digitize the mortgage loan application process, but we have yet to see real adoption of technologies to improve loan processing, underwriting, and closing. Optical character recognition (OCR) and natural language processing technologies that convert paper documents to digital, structured data, and loan origination system software that leverages more modern, efficient workflows will be key enablers.

Jane Mason, CEO & Founder, Clarifire: For mortgage lenders and servicers, no-touch processing is imperative. Mortgage processes need to become seamless with servicing processes. If servicing is your focus, providing borrowers with access to self-service tools and creating customer-centric responsiveness will depend on having advanced portal technology that is interactive, launches automated workflows and sends digital responses. Borrowers receive results in seconds or minutes, rather than hours or days. Offering this type of rapid digital efficiency is a win-win for all as well as the key to both a servicer’s success metrics and end-to-end mortgage industry modernization. It also improves customer retention by increasing opportunities for originators to provide new products and services to the borrower.

Josh Reicher, Chief Digital Officer, Cenlar: The mortgage industry, and especially the mortgage servicing industry, has struggled to stay ahead of both increasing operational expenses, and customer demands for the digital experience. There is a race to automation to address both—to find immediate opportunities for improved effectiveness and reduced risk, as well as preparing for improved scalability and flexibility when the market returns to full strength. Whether through robotic process automation (RPA), document understanding, process mining and decision management, system integration, or workflow automation, automation technology is a central theme to remain competitive.

At Cenlar, we are automating, for example, many of our controls with the goal of reducing manual tasks, freeing up our employees to do more creative and analytical work. There’s the added benefit that automation also helps remove the risk that comes with manual processes. Importantly, automation helps prevent errors, and create a faster, more efficient process for homeowners.

James Vinci, CTO, Selene Finance: In 2,023, several technology enhancements will play an increasing role in making all employees more efficient, ultimately creating opportunities and a competitive edge. As mortgage companies continue to be driven by negative economic factors, efficiency and effectiveness remain key drivers.

It is natural to focus on frontline and boarding automation activities, but usability on manual tasks, gaining insights within data, and back-end activities, like information security and compliance, are also critical.

Selene has critical active initiatives with respect to: RPA in the contact center; artificial intelligence (AI) based document indexing and data extraction for servicing transfers and post-boarding quality reviews; compliance and data collection automation for information security; and internal and external customer digital experience tracking and optimization. We are also always working to improve how our on-premises and remote team members engage and interact.

Mark Walser, President, Incenter Appraisal Management: Lenders need to adopt technologies that enable them to continue streamlining processes, making transactions more affordable, and speeding time to close. For example, as a market recovery occurs and mortgage volume picks up, lenders cannot go back to relying solely on the standard appraisal process, which increased overall turn times and added as much as two additional weeks to the lending cycle during the 2020–2021 housing boom. They will need the capability to fulfill desktop appraisals—using virtual inspections by trained, professional appraisers. Solutions like these also empower lenders to reduce expenses that they might normally pass along to the consumer. With the housing industry seeking to stay strong while making homeownership more accessible to everyone, these technologies are a win for all.

Louis Zitting, Founder & CEO, MonitorBase: Lenders will need to invest in automation and AI-driven technologies to gain insights into the behavior and preferences of potential borrowers. This will allow them to target their marketing efforts more effectively and make more informed decisions about loan approvals. As the mortgage market comes back, business owners operating with a much smaller staff will look to their technology providers to fill the gaps.

 

Is the industry any closer to an all-digital mortgage process? Or do you feel that the inclusion of the human touch during the process will prevent an all-digital process from ever becoming a reality?
Danna: We’re much closer to a fully digital process than ever before, at least in terms of the data used to underwrite and process loans. But digital lending is not an alternative to the human touch. Even with all our digital data, from application to close, we’ll still be counting on knowledgeable loan officers, careful underwriters, accurate appraisers, and personable closing agents to do the work.

Angela Hurst, Chief Administration Officer, USRES Family of Companies: Surveys show that today, borrowers lean towards digitization more often. But when it comes to outside-the-box requirements and critical milestones in the process, personal interaction is preferred.

Additionally, the mortgage process remains extremely complex and heavily regulated during the origination process and well into the life of the loan and servicing. Regulations at the federal, state, and local level continue to dictate one to one contact with the borrower, preventing an “all-digital” process from becoming a reality.

Matt Lehnen, CTO, Deephaven Mortgage: The mortgage industry has made enormous progress towards fully digital loans. Borrowers can shop for products, apply online, eSign documents, link accounts, and close loans without ever meeting face-to-face with their lender. The differentiator for lenders, though, is the human touch, experience, and comfort that borrowers appreciate. Borrowers want the efficiency of digital processes while knowing that there is a real person on the other side who truly cares about making their dream of owning a home a reality. That’s especially true in the non-QM space, where the guidance from expert mortgage brokers and loan officers on different products, features, and terms is invaluable.

Levin: No matter how “digital” the process gets, some borrowers are always going to want to work with a loan officer to help guide and direct them through obtaining a loan and coach them on product options and financial implications. However, the back-end tasks in the process, like underwriting, should become more automated over time as the technology improves, allowing for a largely digital fulfillment process.

Walser: Human touch will continue to be an important component of closing a mortgage. It’s the process that humans use that will change. Take digital closings, for example. Like a contactless appraisal, the idea will be to save paper-processing costs, enable signing anywhere, and preserve the veracity of the closing using technology. There is growing acceptance of processes like these, but change will have to work its way through regulatory and governmental hurdles. My opinion is that lenders should focus less on trying to get to an all-digital mortgage process, and more on “digitizing” the parts of the process where technology is ready to bring efficiency and speed.

Zitting: The industry is constantly getting closer to an all-digital mortgage process, but I still think there is a need for human interaction because the homebuying and mortgage process is so emotionally driven. At the end of the day, people are better suited to guide borrowers through their emotions in order to achieve their financial goals.

What advancements in artificial intelligence (AI) and machine learning (ML) will take the digital mortgage process to the next level?
Danna: We rely heavily on advanced technologies like AI and ML. They are already embedded in some form in most of our systems, and much of the technologies we employ today. Beyond the mechanical functionality, the next steps are more likely to be about human adoption of our advanced tools. For instance, we have appraisal management technology today that is fully automated and provides expansive data and insight into the quality and performance of the lender’s processing team and appraiser panel. Is the lender’s appraisal department using this information? Going to the next level will involve making the most of today’s modern mortgage technology.

Hurst: To have significant advancement it requires partnering with regulators to adapt to today’s regulatory requirements, and a continued partnership to prepare for the evolution in design, creation, and adaptation of technology supporting the digital mortgage space.

Lehnen: Transparency between the machine and customers will be key to adoption and long-term success. With a human conversation, the unique details of a borrower’s situation are discussed first-hand, and questions can be answered in real-time, whereas a machine only has data elements as inputs and a set of decision criteria to reference.

For borrowers with simple and straightforward loan scenarios, full automation may be particularly useful for predictability and transparency. Providing insight to the borrower as to how decisions are reached can help build trust and long-term participation.

In the non-QM space, AI and machine learning can assist by reducing processing and underwriting time. For example, certain technology tools are leveraged to assist in analyzing bank statements, including tools that use AI. Still, the judgment of trained and experienced underwriters will always be necessary behind the scenes. Relying solely on automation can lead to risk models that are overly permissive or restrictive. In the future, advancements in AI and machine learning to offer an even better blend of self-service technology and a seamless experience will take the all-digital mortgage process to the next level.

Rob Nunziata, Co-Founder & CEO, ActiveComply: AI and machine learning can help with compliance-related tasks, such as monitoring your company’s social media or remote workers. Also new technologies are helping to make sure data on files is correct as it passes through different departments. The old adage of “stare and compare” is now being replaced by some AI technology that is way faster and more accurate than humans.

Reicher: Machine learning is taking technologies like virtual agents and automation technology to the next level. Stepping beyond the natural language processing required of modern virtual agents, machine learning can leverage a customer’s mortgage information and activity to personalize the digital experience.

Virtual agents can adapt based on recent activity, both mortgage and customer, to customize and maximize interactions. Machine learning allows automation technology like RPA to adapt and improve over time with less overhead, and fewer exceptions to be worked by human counterparts.

With the advent of AI-based language models like ChatGPT, virtual agents will take on an even broader role in the mortgage process. Virtual agents leveraging these new models will have access to a massively expansive database, coupled with the ability to communicate this information in a human-friendly manner, which will transform customer service.

Souren Sarkar, CMB, President & Co-Founder, Nexval: AI and machine learning algorithms are already being used in the mortgage industry to generate innovative solutions and value for both consumers and financial institutions. Even so, mortgage processes are complicated and may still require several manual inputs for the foreseeable future.

However, two areas where AI and machine learning are helping to advance digital processes are improved application verification and better document validation. With applications, for example, AI-based mortgage IT companies may be able to use a lender’s existing data for predictive model training, which makes it easier to estimate a borrower’s income.

This information could then be compared to the borrower’s reported income to ensure it is accurate and reasonable. Machine learning models may be improved to verify the consistency, thoroughness, accuracy, and validity of data in loan documents, since mortgage financing remains a document-intensive process that involves extensive paperwork to close a loan.

Another future advantage of implementing AI is that loan modifications can also be done as needed. With advancements like these, it’s not an exaggeration to say that AI will be the future of mortgage processes.

Vinci: Advancements in AI and machine learning could potentially advance the all-digital mortgage process in a number of interesting ways. AI and machine learning are already being applied to help automate the underwriting process by quickly and accurately assessing a borrower’s creditworthiness.

This is being extended to origination quality control processes, secondary market due diligence reviews, fraud detection, and servicing transfers. It can greatly impact the accuracy of document indexing, data extraction, facilitating doc-to-data reviews, and other quality control reviews. AI and predictive analysis can also help forecast loan modification recidivism and force majeure event impact models.

Natural language processing, in combination with AI, can be used to decrease customer contacts to deliver requested information, reducing the need for human intervention. Other interesting areas include the use of virtual or augmented reality for property inspections, and for potential buyer showings.