Here’s how mortgage lenders are thinking about AI in 2025

As technological innovation is geared to propel the mortgage industry into a new chapter of efficiency, many players in the space are excited for change, but taking their time with adoption.

Bạn đang xem: Here’s how mortgage lenders are thinking about AI in 2025

There is a desire to incorporate new AI tools to streamline operations going into 2025, but simultaneously there is a holdback by some companies because of financial constraints and regulatory unease.

Bigger players with cash reserves, such as Rocket Mortgage, United Wholesale Mortgage and Guaranteed Rate, have enthusiastically jumped into innovating in-house. 

Others, especially more mid-sized shops, are being more picky about adoption, oftentimes opting to rely on third-party vendors to test AI offerings and being intentional with what tools are being incorporated.

Company leaders say they see artificial intelligence-based tools as a means to cut significant costs in their operations, but it is a long-term investment, which is a hard pill to swallow especially during times of weak origination activity.

Relying on vendors to build out systems can cost anywhere from $20,000 to $200,000 depending on the project, stakeholders have said. Meanwhile, for companies building proprietary technologies, such investments can cost millions of dollars.

Artificial intelligence can be used for lead acquisition, streamlining the origination process and crunching data to find borrowers ready to refinance. But the amount of investment needed to fully tap into AI potential leaves many unable to jump in — yet.

The bigger you are, the easier AI adoption is

In the past year, lenders such as Guild Mortgage, Loandepot, Rocket Mortgage, Union Home Mortgage and others have announced plans to increase their use of AI.

Filings with the Securities and Exchange Commission show that on a quarterly basis, large public nonbanks spent millions of dollars on technology upkeep. These companies are actively deploying AI solutions.

Xem thêm : US mortgage market in 2025 – will it turn the corner?

Better, which recently launched its AI voice assistant Betsy, revealed that its technology expenses totaled $7.2 million in the third quarter, up $0.9 million, or 14%, from $6.3 million in the same period last year. Meanwhile, Pennymac spent $37 million on technology-related expenses.

Some of this tech investment has been enabled by a consistent flow of income from servicing.

According to Bill Cosgrove, CEO of UHM, the company has a “nice size servicing portfolio” which has “saved the day” in making business profitable and as such has given the lender the ability to invest into technology.

Cosgrove said that UHM is investing into AI to help streamline servicing operations, which will pay off in the long term. The company, which has over 100 employees in its IT team, is currently testing out an AI voicebot, which it hopes to launch sometime in 2025.

However, for more mid-sized companies, especially for those that don’t have servicing portfolios to buoy business, AI investments may not be as feasible in the current economic landscape.

Lenders interviewed say interest is there to implement AI technologies to streamline mundane tasks, but most are being prudent with what they invest into.

Synergy One’s CEO Steve Majerus notes that his company is not throwing all company resources into “shiny AI tools,” but is investing enough to help loan officers compete with other lending shops. 

“One of our operating premises for the company is helping our loan officers create modern mortgage experiences for homebuyers and homeowners,” said Majerus. “While we’re being cautious, we are optimistic about what AI holds for improving the customer experience, our customer analytics, and the outcomes that it can provide.”The company’s CEO says the San Diego-based lending shop has strived to be part of the wave of cutting-edge technology adoption, but cautiously and with the help of vendors.

“Our focus is to grab the low-hanging fruit where AI can provide value to a company like ours in areas like database management and retention of our customer database, and provide analytics around opportunities and managing that database for our loan officers,” said Majerus. 

Jeff Bode, CEO of Clink n’ Close, said his company is “taking bite size pieces” of implementing AI into its workflow. Investment thus far has been into using AI to “understand who we’re marketing to better and how best to reach them.” There are plans down the road to invest into an AI chatbot, but there is no immediate rush for now.

There is an overall sentiment in the industry that investment into AI is worthwhile. A survey by Arizent, parent company of National Mortgage News, found that 47% of mortgage professionals surveyed say updating technologies is a priority going into 2025, up from 39% the year prior.

Other constraints and considerations

Xem thêm : Better time to buy than refinance

The obvious constraint among most mortgage lenders to integrate new tech tools is the cashflow problem, but also a priority issue, Bill Dallas, former head of Finance of America and current consultant, said.

“Most of my smaller clients have PTSD with respect to interest rates and losses, so AI investment is not even on the radar for 2025 because it’s all about survival,” he said. “You have to have sizable capital in order to actually try to attack this.” 

Dallas noted that instead of prioritizing tech investments, most smaller shops are trying to “figure out where the next dollar is coming from,” which will lead lenders to hire originators rather than improving internal processes.

“[Lenders are thinking about] ‘do I have another loan and could I bring on another branch and steal a bunch of people from Movement Mortgage or Guaranteed Rate.’ They don’t think about fixing their broken processes first,” Dallas said.

Gabriel Skelton, head of artificial intelligence solutions at OpenBots, added that some lenders who could benefit from AI are hesitant to invest due to past negative experiences with subpar products from vendor partners.

“Mortgage lenders have been targeted by technology companies a lot and that has left a bad taste in their mouth because the offerings haven’t been that great,” Skelton said. “It takes a certain level of innovation and validation of what you’re working on to really convince lenders.”

It is also hard for some to pin down exactly how much investing into AI tools and gadgets will save companies in the long term.

“I think it’s too early to determine the return on the investment,” said Cosgrove. “But I do think it’s a matter of speed and accurate information coming from the machine learning or the bot.”

A Freddie Mac survey on loan production costs found a significant gap between the top and bottom 25% of lenders. The bottom 25% spent $16,500 to produce a loan, while the top 25% spent $6,900. Skelton argues mortgage shops investing into AI are the ones tapping into those savings.

“Every lender has one cost to manufacture per loan based on all the people that are touching the documents and touching all the data. The more people you have, the higher the cost to produce, while the more efficient you are, the lower your costs will be,” Skelton said. ” Even if technology can be pricey, it adds value.” 

For lenders that haven’t invested, but want to invest into AI, Dallas recommends “being a fast follower.” 

“Companies need to look at the bigger mortgage companies and maybe what big tech lenders invested into in other industries to see what you should be investing into. I don’t think this is one that you’re going to get a quick return. This is long term.”