Premier Member Editorial: The Next Mortgage Cycle Belongs to Intelligent Automation, Not AI Hype

Casey Logan Williams is General Manager, nCino Mortgage

Early hockey games were played with improvised pucks—wooden blocks, even frozen lumps of material, none of which survived long on the ice. The modern puck only became possible once manufacturers in Canada turned to vulcanized rubber, a material that could withstand repeated impact in subzero conditions without splintering.

The mortgage industry is facing its own version of that shift. AI dominates the agenda at every major conference, and lenders are feeling immense pressure to adopt something, anything, to keep up. Rising operational costs, shrinking margins and growing borrower expectations create urgency, but progress doesn’t come from chasing whatever appears most advanced. It’s about adopting technology that genuinely streamlines workflows and cuts through the friction that slows lenders down day after day—and that’s where intelligent automation (IA) comes in.

Where intelligent automation transforms lending

Basic automation has supported lending for years, but its impact is limited to task execution, such as pulling credit. IA, by contrast, interprets and acts on information—analyzing the liabilities, determining the credit score and offering a decision on the borrower. Instead of just automating a task for borrowers, such as pulling W-2s or paystubs, it will analyze those diverse income types, extract the data, flag potential discrepancies, and then predict the likelihood of loan approval based on industry standards, or even pass that information back to the underwriting system.

According to the Freddie Mac 2024 Cost to Originate Study, origination costs have risen nearly 35% in just three years, adding roughly $3,000 per loan. As a result, personnel expenses, lenders’ largest cost, are being squeezed. Fannie Mae’s 2025 Mortgage Lender Sentiment Survey reports that 59% of lenders plan to cut expenses by reducing back-office staff. Many loan officers and support teams spend hours daily answering predictable borrower questions, collecting documents, entering data, and reviewing files—the very operational burdens IA is designed to reduce.

Although IA often incorporates AI or machine learning, it doesn’t depend on them. AI is just one tool among many. What matters is whether a solution addresses operational gaps that slow down borrowers and staff. Lenders see the greatest impact when they begin by examining where work breaks down—repetitive steps, preventable delays, predictable questions and points of fallout where manual intervention becomes routine. These high-friction moments, especially in high-volume stages of the mortgage process, create the strongest opportunities for early ROI.

Knowing whether to build versus buy

Once lenders identify where IA can help, they face a critical decision: build in-house or partner with a provider. While many lenders explore internal development, most institutions are built to excel at lending—not at running long-term software product organizations. Lenders must determine whether diverting teams and capital toward platform development meaningfully advances the core mission of closing more loans, faster.

Internal builds can take years to reach maturity, and by the time they do, market needs or operational realities may have already shifted. Meanwhile, the cost of ownership—supporting, scaling, and continuously updating those systems—can quickly eclipse early expectations. Even a technically strong solution may struggle to gain adoption if it arrives too late or lacks alignment with day-to-day workflow demands.

Just as important is the foundation on which IA operates: data. IA is only as effective as the data feeding it. Lenders need clean, consistent, well-structured and accessible data to ensure IA can interpret pipeline activity accurately. Many organizations believe their data is strong yet lack the outside perspective needed to uncover systemic inefficiencies or blind spots. The right partner can help evaluate data readiness, identify bottlenecks and provide industry context that strengthens IA strategy.

Choose partners that drive impact

Choosing the right partner requires looking beyond polished demos to understand whether a provider has the stability, data maturity, scalability and implementation support needed to make IA truly effective. While emerging vendors may offer clever ideas, many lack the depth required for enterprise-grade deployment. Sustainable IA depends on a partner who can evolve with a lender’s needs—surfacing efficiency gaps, providing best-practice guidance and supporting long-term operationalization.

Measuring impact from the start is equally essential. Many lenders deploy automation without first establishing baseline metrics, making it difficult to evaluate whether cycle times shortened, manual work decreased or borrower experiences improved. Without “before” data, ROI becomes ambiguous and momentum stalls. Defining early metrics—such as cycle times, manual touches, borrower satisfaction and cost-per-loan—creates a clear framework for assessing progress. Once baselines are set, improvements become visible and measurable, guiding leaders on where to expand IA next.

As AI noise grows louder, lenders may feel pressure to adopt tools that appear sophisticated but offer little real value. Without a clear understanding of the problems they’re trying to solve, even advanced technology can quickly become an expensive distraction. Success will come from choosing solutions that help teams work smarter, improve efficiency and strengthen their ability to compete in a demanding market.

(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.)