How Can AI Actually Help the Mortgage Industry? Annual24 Panel Weighs In

(From left: David Tepoorten, Olivia Peterson and Laura Escobar)

DENVER–It’s no secret that artificial intelligence is an increasingly hot topic, and there are real applications to consider for those working in mortgage banking and related industries, said experts from Amazon Web Services and NVIDIA Corp., at MBA’s Annual Convention & Expo Oct. 29.

The panel, “The Promise and Possibility of AI,” was part of Annual24’s Tech Elevated track, providing more than 12 hours of tech-focused content.

To start, Olivia Peterson, Director of Worldwide Public Sector Financial Services for Amazon Web Services, took a few minutes to explain what, specifically, generative artificial intelligence is.

While artificial intelligence takes data and can mimic decision-making processes of humans, generative AI creates new content, Peterson said. “Generative AI actually takes that data, continually learns off of past recommendations, as well as new training data that we feed it, but actually creates what new content might look like. So that is really the next generation of artificial intelligence,” she said.

In terms of use cases, panel moderator and 2025 MBA Chair Laura Escobar, also President of Lennar Mortgage, asked about natural disasters. Hurricanes Helene and Milton were a common topic of discussion across multiple sessions at Annual24, including as they relate to the availability and affordability of home insurance.

Generative AI technology may be able to help there, the panelists said. David Tepoorten, Director of Global Systems Integrator Partnerships at NVIDIA Corp., showed an example of Nvidia’s Earth-2 Platform for Climate Change Modeling.

“It is a digital twin and it simulates the Earth. It simulates all the weather conditions, all the precipitation that’s in the air, all the droughts,” Tepoorten explained. “It has all built into this one model that scientists around the world are using to forecast for better accuracy for these types of events.”

Imagine if you could track or predict a hurricane on a much longer timeline than the current 10 days or so we get from traditional meteorology, Tepoorten posited–“If you could track hurricanes, what would that mean to the underwriter? And what would that mean to insurance carriers?”

As for other uses in the mortgage industry, Peterson pointed to AI’s ability to ingest and analyze data, allowing companies to make better and more efficient decisions.

And, some companies in the space are already using intelligent document processing, she said, which has enabled them to go from hours of review for loan documents to minutes.

Peterson queued a video to explain intelligent document processing–it showed how the AI can recognize and classify a set of documents, figure out if they’re complete, and then analyze or provide the data and information included in them in various forms, among other uses.

Tepoorten also pointed to the possibility of AI support for those working on the customer-facing side, noting that it can help employees with decisions such as next steps while on client calls.

“You really have an assistant to the human, which helps get you to the highly personalized view of what that customer may be eligible for, what they may have questions or concerns about, or just things that they may not be aware of, programs that you can bring to their attention,” Peterson described.

However, “I think it’s really important to look at the fact that this is assistive technology,” she noted. “This is something that makes a recommendation, but particularly at these phases where the models are still maturing, where comfort and trust is still being gained, it’s important to have that other point of validation to say: ‘Given these inputs, does this recommendation make sense?’ Because it’s really up to the user to decide [how to] act on that recommendation.”

Peterson and Tepoorten also described digital avatars that can help customers with certain issues, translation capabilities and the identification of fraud or other quality issues as AI use cases potentially relevant to the mortgage industry.

So, what should companies in the space be thinking about AI moving forward?

“I think my advice to you is just to try it out,” Tepoorten said.

Peterson agreed. “The only way this becomes real to you for your use case is to start to experiment,” she said. “So, one thing that is a call to action that I encourage each of you to think about is, what is one use case that either is going to drive efficiencies internally, is going to benefit your customers or mitigate risk in your business that you can work with?”