Bill Packer, COO of leading reverse mortgage lender and servicer Longbridge Financial, previously explained for RMD how generative advanced algorithms widely referred to as “artificial intelligence” (AI) can be used to distribute reverse mortgage product information and process documents. Still, there are other potential uses the emerging technology could be helpful.
Because of the advanced language capabilities of these kinds of systems, there is a unique opportunity to break through language and product complexity barriers to assist in both educational initiatives as well as loan manufacturing processes. To get a better understanding of how these could work, RMD spoke further with Packer.
Reverse mortgage language, loan manufacturing uses
Specifically in terms of short-term applicability to the industry, Packer sees two primary uses: powering up the ability for customers and employees to locate the right information for their needs and streamlining the complexity of a loan’s manufacturing processes.
“The first is really what you might think of as ‘search on steroids,’” Packer said. “The nice thing about these generative AI technologies is you can ask it questions in a variety of ways. Syntax really doesn’t matter, because it has a good understanding of syntax. Spelling also usually doesn’t matter.”
After experimenting with one of the potential AI tools Longbridge has in development, Packer misspelled the word “farm” by typing in “F-R-A-M.” The generative tool didn’t miss a beat and understood what the input meant.
“It gave me the right answer,” he said. “It knew that I meant ‘farm,’ and you can ask it across a variety of languages. In Spanish, for example, or in a variety of Chinese dialects. And it understands that, so you can imagine that we are able to expand the universe of how we can assist both our own employees and other folks in understanding the reverse mortgage product.”
There could also be an opportunity to help streamline the loan manufacturing process with this technology, he added.
“The other side of this is the extent to which we can develop aids that streamline this process, and help people ask the right questions,” he explained. “[We envision a tool to help customers] obtain the right documentation, structure the loans correctly and easily optimize for the customer’s goals. I think that’s another place where we’re going to see a lot of a lot of use here.”
Applicable to mortgage companies with an online presence is the potential ability to automate tech-related tasks, he said.
“I see a lot of companies using it for routine tasks, like password resets,” he said. “So, I think that’s another place where we’ll begin to see some of this technology get introduced. Those, to me, are the critical pieces.”
AI tools are not a replacement for people
Packer added that many companies have seen generative AI technology as a way to transition customer interaction tasks to chatbots, but Packer said that is not in Longbridge’s plans.
“[Some companies may see this tech as a way] to reduce the number of personnel they need to have real-time conversations, but that is not our approach,” he said. “We really see these as assistive, and helping to deliver better information rather than — in any way — replacing that personalized approach. That’s a key piece for us.”
It is also true that as advanced as these bots are, their limits can become very visible depending on the task that is asked of them, Packer added.
“I’ve certainly had the experience where I have become frustrated because they’re not able to deliver the service you need them to,” he explained. “That is definitely not the road we’re going down. The road we’re going down is really one to make it more assistive and as an augmentation to the already great experts that we have on staff.”
Still, Packer suspects that there are some companies in the space that may try to automate high-touch customer interaction processes with generative AI too quickly, but in the reverse mortgage space especially that would be a mistake in his estimation.
“How do you cost justify the investment?,” Packer asked rhetorically. “Do you sit in a boardroom and pencil out the ROI? One of the ways some companies have done that is to commit to reducing people costs, but that would be a mistake. Considering both the product’s complexity and the demographic we typically interact with, we need to have that higher touch. So some may try, but it certainly won’t be Longbridge.”
