One year after speaking at the second annual HousingWire AI Summit, Ryan Grant, president of NEO Home Loans, returned to the stage to share updates on his company’s partnership with Better Mortgage and its technological prowess.
The conversation came just a week after a pivotal leadership shakeup at Better Home & Finance Holding Co., with CEO Vishal Garg stepping down and being succeeded by board member Daniel Lewis as interim CEO. NEO has been partnered and powered by Better since January 2025.
Sitting down with HousingWire CEO Clayton Collins at the third annual AI Summit on Tuesday, Grant said he views the change as a natural evolution. It comes as Better reaches a point where it needs to focus on executing the technology platform it has built before pursuing its next wave of innovation, he said.
“Every company in general kind of reaches a point where the vision takes you so far,” Grant said. “Without that operator and without the disciplined execution … the future stuff becomes noise.”
The transition has raised questions about whether Better will continue investing heavily in artificial intelligence and other technology amid concerns about profitability and spending, a topic that Collins asked Grant about on stage.
“We have barely scratched the surface,” Grant told Collins. “We’re trying to figure out how to move from an antiquated, human-centered industry … to what is possible if we get this technology right.”
Calculating the cost to produce loans
Aside from addressing Better‘s leadership shuffle, much of Grant’s discussion with Collins centered on how AI is affecting the cost of producing a mortgage.
Grant argued that the industry first needs to establish a more standardized definition of “cost to produce,” noting that lenders calculate the metric differently depending on whether they operate in the retail, wholesale or direct-to-consumer channels and which expenses they include.
At NEO, for example, Grant said the company focuses on the mortgage manufacturing process, including underwriting, closing, funding, post-closing and shipping. Grant said that underwriters who previously handled two to three loans per day can now process eight to 10, with about half of these files requiring little or no manual intervention from the underwriter.
Closers and funders, meanwhile, have increased their productivity by roughly 90%, Grant said. He estimated NEO’s cost of production at roughly $700 to $800 per loan in its direct-to-consumer operation and about $1,200 to $1,300 in retail. The figures exclude sales and origination commissions.
The retail figure is higher, he said, because NEO handles a broader range of products, including non-QM and other purchase products, while its direct-to-consumer business has a narrower credit box.
As AI is changing the mortgage workforce, Grant said that despite productivity gains, NEO can’t simply eliminate staff, Grant said. The company has grown by about 100% over the past 18 months, driven by new employees and increased production from existing teams.
Grant said that NEO aims to operate at about 75% staffing capacity, leaving roughly 25% of capacity available if mortgage volumes increase. Grant added that NEO has moved away from a traditional processor role as technology takes over more of the administrative work, leading preapproval specialists and loan specialists to become focused more on the consumer experience.
The goal, Grant said, is to allow mortgage professionals to “go be human” with consumers rather than spending their time managing processes and paperwork.
The same philosophy applies to originators. Grant said the best mortgage professionals should not need to spend their days inside a customer relationship management system or loan processing platform. Instead, technology should operate around them while they focus on business development, education and client engagement.
Winning consumer trust in AI
Getting employees to trust AI remains one of the biggest challenges to adoption. Using an interaction with a recently hired underwriter who continued to manually review loans that had already been underwritten by Better’s AI system as an example, Grant said that the underwriter wanted to make sure the loans were correct because his name and license were attached to them.
Grant said the conversation illustrated the cultural hurdle facing mortgage companies that deploy AI. Employees need to see the technology as a partner rather than a threat to their jobs.
NEO has addressed some of that resistance by making its technology the primary system employees use rather than offering it as an optional tool. That has resulted in full adoption, Grant said, although employees are still learning how to take advantage of the technology’s more advanced capabilities.
“They just need to know you can trust it,” he said.
This article was written by Sarah Wolak and generated with the assistance of HousingWire Automation, then reviewed by a HousingWire editor before publication.
