UWM, the top U.S. mortgage lender in the first quarter of 2024 with $23.4 billion in volume, has developed its technology in-house with “very few partners,” a method it used to “control the product,” Bressler said.
Bressler said that UWM gives its technology away for free so brokers can compete against anybody, creating loyalty to the lender. Its technology also provides scale to navigate the ups and downs of the mortgage market.
Bressler and the other executives spoke during the MISMO workshop on AI and mortgages held in San Francisco, California, on Monday.
Dan Vasquez, who leads the AI strategy at Rocket Mortgage, said the Detroit-based company believes that AI is a “great tool,” but it “doesn’t absolve you from the responsibility that you already hold in making decisions that impact the client.”
“It might give you tips, it might give you new information, it might give that all to you faster, but it’s still your responsibility to make the right decision. And we’re not taking that away from humans right now,” Vazquez said.
In April, Rocket launched its patented AI-driven platform, Rocket Logic. It also added Synopsis, an AI tool that allows its client experience teams to analyze sentiment, record behavior patterns, transcribe conversations, and create tailored experiences during customer interactions. Rocket is the third-largest mortgage lender in the U.S., according to Inside Mortgage Finance, with about $19 billion in production from January to March.
According to Vasquez, two-thirds of income underwrites at Rocket don’t “touch a human” as the lender uses AI models to process documents and verify the information with a high degree of certainty.
“I wouldn’t say that there are plenty of places where AI is not a good fit yet because we haven’t yet invested in the right amount of models, data, and techniques to make it work. We’ve got a pretty proven framework for how to do so. But it takes a lot of human power. It takes a lot of time to train those models to curate the right data to move forward into different levels of underwriting or decision-making to do it safely. I think we’ll get there in almost all places, but I don’t think, at least not in the next couple of years, we’ll get to a place where we decide: ‘Okay, we humans don’t need to touch these loans anymore.’”
Vasquez added: “We will definitely run into places where we think the technology is doing better than the humans were before. And it’s okay to peel back a little bit. But you always want to make sure that you have that [human] backstop.”
Josh Friend, chief executive officer and founder at Insellerate, a customer relationship management and engagement platform, said that despite growing AI adoption, loan officers are not using it to automate communication.
“We’re making it so that the loan officer has to choose whether or not they’re gonna send that email out as of today, the reason being is you have to watch AI,” Friend said. “I think at some point we’ll be able to not have to do that level [of oversight] because of enough learning, enough practice that will take place. But, for now, we don’t want the AI to go ahead and make the decisions for us without some human being able to have some oversight with it.”
Kiran Sahota, executive vice president of technology at Freedom Mortgage, said that AI or generative AI isn’t something new but can transform mortgage businesses and the industry overall.
“We don’t think we’ll eliminate the human loop. It’s just we’ve got to rescale them and give them the superpowers to be able to do the job faster, cheaper, better, so we don’t have to go to hiring and firing cycles,” Sahota added. “We are in a regulated industry, so we need to do this responsibly and ethically.”
Freedom has rapidly grown, becoming the fourth-largest mortgage lender in the country from January to March, with a production of $11.8 billion in mortgages in the period.
Sahota said Freedom has also grown its servicing portfolio fast to about $507 billion in unpaid principal balance in the first quarter. The lender’s tech strategy has shifted from buying to building over the last four years, a decision guided by factors such as urgency and differentiation.
