Eastern Union's Abe Bergman says his team now knows who's active and whose debt is coming due
The rapid increase in artificial intelligence development has spread across the mortgage industry. In the often complex world of commercial mortgage lending, AI is providing both the ability to expedite once tedious tasks and a wealth of market information at a broker’s fingertips.
With fewer deals closing, commercial brokers can't wait for business to come to them, and one firm is using AI to find the borrowers who are about to need a loan.
At Eastern Union, AI now touches nearly every step of the mortgage journey, from lead generation to loan processing.
Abe Bergman (pictured top), president of Eastern Union, said AI has even changed how his team finds clients.
"The quality of my prospecting has 10x'd because I know I'm reaching people with the right number,” Bergman told Mortgage Professional America. “I know I'm reaching people that are active. I know I'm reaching somebody who has a loan coming due. The quality of my outreach has gone up, and we've seen it translate in very short order into more potential business.”
Finding loans before they mature
Eastern Union first used AI to underwrite faster and respond to borrowers sooner, Bergman said. The prospecting push is more recent.
"Over the last, I'd say, 6 to 8 weeks, we've really fully incorporated this into our lead generation, our business development side of it," he said. "Really being able to identify which clients need help right now. Who has a loan maturing? Who has a loan that potentially needs to be refinanced? Who's buying? Who's active? Who's not active? And then being able to provide our salespeople, our brokers, the ability to very efficiently be able to do outreach to them on a very personalized basis."
According to a survey released in February by the Mortgage Bankers Association, about 17% of the $5.0 trillion in outstanding commercial mortgages, or $875 billion, were set to come due in 2026.
AI also changed how the firm tracks who is making loans. That research has grown heavier over the past month or two.
"The underlying foundation of a good mortgage broker is his or her banking data. Who's lending on what?" Bergman said. "We're in a market today where lending is changing every day. And today, with the immediate access to data, we know immediately who closed what. And now we can very easily say, ‘I have this kind of deal. Who in the market in the last 3 months closed the deal?’ Because if it's 6 months ago, it's old data today."
Eastern Union has an in-house programmer and works with two outside AI consulting firms to help develop tools. Bergman said most of the ideas start with the staff using the tools.
"A huge part of our adoption over here has been because we have people in the office who pop up and they say, ‘Hey, Abe, I just tried this and look what it did for me,’" he said. "And then somebody else says, ‘Oh, you know what, why don't you tweak it that way?’"
Managing rapidly changing tech
Getting the data right is the first concern, Bergman said. The firm has set limits on where its tools can look.
"Banking data is only from sources that we allow it," he said. "So we'll restrict where it actually searches its data. It's internal data, certain public data that we purchase that we know is relatively accurate, a high degree of accuracy. Trust and verify hasn't changed."
Brokers also can't count on AI to replace their paid data sources, he said. Bergman used a search for rental comps as an example.
"You can't just say, ‘Okay, I'm going to cancel all my subscriptions to data because I can get everything from AI,’" Bergman said. "You're not going to get a complete set of data. So you have to realize that just because you query AI, whichever agent you're using, does not mean that it's 100% complete. Even if it's 100% accurate, it doesn't mean it's 100% complete."
Tools are also changing faster than the firm can plan around them. Bergman said that has already changed what the firm builds.
"We've seen where we put on the schedule tools we want to have the developers develop over the last few months," he said. "And countless times, by the time it got to that, 3, 4, 5 weeks later, we got, ‘Okay, we're ready for this.’ We don't need to develop it anymore. It's been done for us."
Cost is the other open issue for the firm. It pays for some tools and uses others at no charge. The question everyone is wondering about AI tech is how much users will end up paying for it in the future.
"Is that going to stay free? Is the pricing model going to change? Are the ones we're paying for going to get cheaper, get more expensive?" Bergman said. "That's a big, big unknown that none of us users really know the answer to."
Stay updated with the freshest mortgage news. Get exclusive interviews, breaking news, and industry events in your inbox, and always be the first to know by subscribing to our FREE daily newsletter.
