At the New England Mortgage Expo, artificial intelligence wasn’t treated as a concept or a shiny vendor demo. It was discussed as infrastructure, already deployed, shaping workflows, and forcing uncomfortable questions about where human value lies within the mortgage process.
Across panels and hallway conversations, one theme kept surfacing: AI is no longer about replacing people wholesale. It’s about compressing time, stripping friction out of the sales process, and exposing which parts of the business are truly differentiated, and which are becoming commodities.
AI Moves To The Front Of The Deal
As AI tools move deeper into the front end of the loan process, some originators are also looking ahead to how those efficiencies could reshape entire transaction types — particularly refinances — where speed and standardization already dominate borrower expectations.
Scott Valins, co-founder and CEO of Go Rascal, a mortgage brokerage, said the most immediate impact of AI is not underwriting or post-submission operations. Instead, it is showing up in the earliest moments of borrower engagement, when trust is built or lost.
Because Go Rascal operates as a brokerage, Valins said the firm’s focus is on sales-side efficiency. AI tools are being used to reduce administrative work during pre-approvals and loan structuring so loan officers can spend more time on relationships and lead generation. Call transcription, automated note organization, and application pre-population are increasingly becoming standard expectations rather than optional tools.
The underlying assumption is straightforward. Borrowers do not want to repeat themselves, and loan officers do not want to spend time managing paperwork. As technology becomes less visible, the interaction itself can feel more human, even as AI handles much of the background work.
AI As A Delivery System For Expertise
James Jin, CEO and president of General Mortgage Capital Corporation, described AI less as automation and more as a way to surface expertise. He said the technology helps make complex programs more understandable and accessible to both borrowers and originators.
With hundreds of niche and non-agency programs available, Jin said the challenge is not product availability but accessibility. AI tools can surface complex options more quickly and transparently, allowing borrowers and loan officers to explore scenarios that might otherwise require deep institutional knowledge or multiple rounds of back-and-forth communication.
Jin said this is where AI creates leverage rather than disruption. Simple agency loans are already trending toward standardization. The real value for originators, he said, lies in navigating complexity, including high-net-worth borrowers, asset-based qualification, self-employed clients and scenarios that do not fit neatly into automated systems.
Used correctly, Jin said, AI does not replace expertise. It delivers it faster and to more people.
From “AI-powered” To Operationally Proven
While many companies are racing to attach large language models to mortgage workflows, Lukas Rosenblum, chief client officer at AngelAI, said that approach can introduce risk if it is not grounded in operational consistency.
Rosenblum drew a distinction between probabilistic AI, which generates likely answers, and deterministic systems designed to produce the same outcome every time. In mortgage lending, he said, consistency matters more than creativity. Predictable outputs are what hold up under audits, repurchase scrutiny, and regulatory oversight.
Angel AI positions its platform not as an assistant, but as a transactional system embedded across the full loan lifecycle. Rosenblum said the difference between experimentation and maturity is whether a system can operate at scale within regulatory guardrails.
That maturity, he added, also affects how companies think about staffing and job roles.
Rosenblum said AI adoption should be viewed as a gradual reallocation of time rather than a sudden disruption. Over time, companies can move staff away from repetitive tasks and toward higher-touch borrower interactions.
Disclosure, Trust And The Human line
Rick Roque, corporate vice president of growth at NFM Lending, said the conversation around AI should focus not only on capability, but on deployment.
As AI-powered voice and outreach tools become more common, Roque said transparency is both an ethical and strategic consideration. Consumers tend to respond better when they understand how technology is being used on their behalf, and attempts to disguise AI as a human voice risk undermining trust.
Roque also described AI adoption as a competitive inevitability. Local presence, once a defining advantage for loan officers, is being eroded by data aggregation, automation, and social distribution. Over the next 12 to 24 months, he said, lenders that move quickly will have an advantage, not only in adopting tools, but in integrating them across sales, operations, and borrower experience.
Pressure Reshapes The Market
Taken together, discussions at NEME reflected an industry under structural pressure. Enterprise lenders are using AI to scale faster and operate more efficiently. Brokers and non-delegated correspondent lenders are gaining ground by pairing flexibility with technology-enabled efficiency. Mid-sized lenders, caught between those models, face increasingly difficult strategic choices.
Purchase transactions continue to involve multiple parties, variable timelines, and borrower decisions that extend beyond rate and payment. Refinance transactions, by comparison, are more often driven by defined financial thresholds and repeatable processes, conditions that lend themselves more readily to automation.
Several speakers said that as consumers grow more accustomed to interacting with AI across financial services, refinance activity may increasingly resemble other self-directed transactions, particularly when borrower profiles and loan structures are straightforward. In those cases, technology’s role shifts away from advice and toward execution.
The implication, as discussed at NEME, is not the replacement of originators, but a narrowing of where human involvement is most critical. Complex purchases and nonstandard borrower scenarios are likely to continue requiring hands-on expertise, while standardized refinance activity may place greater emphasis on speed, pricing, and process efficiency.
AI, as framed at the conference, is accelerating that distinction, shaping how different segments of the mortgage market evolve, rather than pushing the industry in a single direction.
