By Rod J. Alba
The mortgage lending landscape is being fundamentally reshaped. Artificial intelligence no longer is an emerging trend. It is a dominant operational force.
AI technology is rapidly and relentlessly injecting itself into every stage of the mortgage lifecycle, from initial loan origination to servicing. This is not incremental change; the technology is advancing by leaps and bounds, driving a profound and immediate need for awareness and adaptation. At minimum, banks are rigorously evaluating AI tools in testing environments to gauge viability.
The financial markets are seeing the development of AI-powered financial tools that claim to navigate every stage of the residential finance process. The tools range from initial customer interface to loan repayment and everything in between.
And adoption is increasing. According to a 2025 survey by Stratmor Group, a mortgage consulting firm, 38% of mortgage lenders in 2024 reported using artificial intelligence and machine learning, up from 15% in 2023. The survey also noted that 48% of lenders used robotic process automation, or “bots,” in the past year to streamline processes such as ordering appraisals and credit scores — up from 30% in 2020.
“Mortgage lenders are increasingly investing in automation and AI as foundational technologies to improve efficiency and productivity,” says Nicole Yung, a senior partner at Stratmor Group. “This represents a fundamental transformation in how mortgage operations are being conducted.”
Yung also says that increased investment in AI includes a “growing trend” of lenders building internal capabilities to “develop and maintain their own bots,” with 21% of lenders reporting that they are developing internal AI capabilities.
“Lenders are strategically prioritizing front-end digital capabilities that directly impact the application process,” she notes. “The industry is at a critical juncture where investments in technology foundations will determine which organizations thrive in the increasingly digital mortgage landscape.”
Some examples of the services and capabilities AI has provided, include:
- AI agents created to guide borrowers through the mortgage application process, automatically extracting and populating data (e.g., income, employment, credit) into forms like the Uniform Residential Loan Application.
- AI programs simulating and processing human conversation, either written or verbal, allowing consumers to interact with digital representatives, often called “chatbots,” as if they were communicating with a real person.
- Mortgage document processing technologies that can be used to handle paperwork such as bank statements, tax forms, income verifications and other documents that need to be considered, summarized and retained in the mortgage approval process.
- AI solutions that can detect fraud and financial crimes by analyzing large datasets for anomalies, verifying document authenticity and identifying suspicious patterns in borrower and transaction behavior in real time.
- AI tools at the mortgage settlement table that are automating closing package reviews, fee reconciliation and compliance checks, and often cutting processing times significantly.
- Mortgage servicing operations automated by AI functions that handle customer interactions, predict delinquencies and manage payment plans through intelligent agents that provide real-time support and personalized outreach.
As AI capabilities advance, mortgage banking executives are increasingly concerned about whether they are moving fast enough to align with the accelerating AI revolution, and many fear that delays could leave them at a competitive disadvantage. These concerns are well founded, but the real compass that should guide bank leaders is that while it is true that the AI revolution is already underway, its transformative impact is only beginning — we are only starting to understand the contours of an AI future where real estate lending operations will evolve in diverse ways.
Commencing the AI “voyage” is the single most important action a mortgage lender can take right now. As Gabe Minton, EVP and chief information officer at Mortgage Connect, advises in a recent American Bankers Association webinar, AI’s power is not magic — it is a deeply potent, results-driven tool that must be integrated into your operations. The goal is momentum, not perfection. Managers must reject the notion of an “all-or-nothing” commitment. The immediate focus should be on exploring targeted AI solutions, identifying specific applications and strategically weaving the technology into the fabric of the bank’s systems.
As banks begin this journey, leadership must focus on what many experts have coined as the “three pillars” of responsible AI adoption: risk management, governance framework and security and compliance. In terms of risk management, executives must remember that AI technology introduces new risks into banking operations; these risks include model bias, inaccurate predictions, operational failures and reputational harm. The governance pillar refers to AI decisions being transparent, explainable and aligned with regulatory and ethical standards. In addition, there must be clearly defined roles and responsibilities for AI oversight across business, compliance and technology teams. The third pillar is security and compliance, which recognizes that AI systems must handle sensitive borrower data, making cybersecurity and regulatory compliance critical and requiring enforcement of data privacy protocols (the Gramm-Leach-Bliley Act in the U.S., the E.U.’s General Data Protection Regulation, the California Consumer Privacy Act and U.S. fair lending laws).
The importance of these three pillars is perhaps best explained by describing their interaction. When taken together, they create a hard shell of trust: governance sets the rules and apportions roles, security enforces them, and risk management monitors outcomes — ensuring AI adoption is responsible, resilient and effective.
In addition, banks should understand that there currently are no federal AI-specific regulations for banks. However, transactions assisted by AI are subject to existing consumer protection laws such as the Truth in Lending Act and Equal Credit Opportunity Act and frameworks like SR 11-7 and the National Institute of Standards and Technology’s AI Risk Management Framework to evaluate AI use. SR 11-7 is a foundational supervisory guidance on model risk management issued jointly by the Federal Reserve and the Office of the Comptroller of the Currency in April 2011 (the Federal Deposit Insurance Corporation later adopted it in 2017).
While written 15 years ago, SR 11-7 remains the primary framework used to examine AI and machine learning in banking. Regulators use their principles to ensure that “black box” algorithms are transparent, fair and free from bias. Likewise, NIST’s AI Risk RMF is a voluntary, customizable non-sector-specific guide developed to help organizations manage risks associated with AI. Released in early 2023, it is considered a “gold standard” for trustworthy AI governance because it provides a flexible, structured way to identify and mitigate AI risks across the entire system lifecycle, from initial design to decommissioning. It includes a generative AI profile as a companion document.
“AI adoption in mortgage lending is inevitable; unmanaged AI risk is not,” observed Eric Lapin, managing partner at Finfusion Consulting. “As AI reshapes the entire mortgage lifecycle, leadership must ensure it is deployed responsibly. It must be governed transparently, secured rigorously and managed with the same discipline we apply to credit, capital and consumer protection.”
Rod Alba is ABA’s SVP for real estate finance.
ABA can assist banks as they employ this quickly evolving banking technology. Visit www.aba.com/AI, which features links to a variety of resources such as the model risk management guidance and NIST AI RMF mentioned in the article, in addition to news, advocacy materials and more.
