Editor in Chief Sarah Wheeler sat down with Melissa Langdale, president and chief operating officer of The Mortgage Collaborative, to talk about the recent FHFA TechSprint focused on gen AI use cases in housing. The use case developed by Langdale and her team was recognized as the most promising use case of gen AI for consumer experience.

FHFA’s second annual tech sprint featured 12 teams that presented use cases in four areas: consumer experience, assessing creditworthiness, operations and risk management, and compliance. Langdale’s team proposed an app that would deliver an Uber-like experience for multifamily renters. Langdale has been in the mortgage industry for more than 20 years.

Sarah Wheeler: A tech sprint always requires a lot of execution in a short amount of time. How did your team decide on what they wanted to develop?

Melissa Langdale: The problem statement we picked was around solving for consumer experience, and our group happened to be multifamily. My entire background has been on mortgage and single-family with some exposure to commercial underwriting and multifamily, but very limited. Our team was very diverse in experience levels and background, and I think that helped us come up with a good solution, which was around consumer experience, tenant screening and some credit and underwriting.

SW: What were some of the things your team considered?

ML: Well, mortgage has a very streamlined process that all consumers go through, whereas in the multifamily side, it’s very segmented and there’s not really consistency from a consumer experience perspective. There are some search sites where they can see rent ranges and availability, but there isn’t a consistency between credit criteria, for instance. So we were trying to create an opportunity for the consumer to be in control of their experience and give tools for standardization of the tenant screening process that the landlord could then leverage to create operational efficiencies. It was really geared at creating some standardization for the industry, putting the consumer in control of it, and then building in a significant benefit to the landlords on the back end.

SW: How do you think your experience on the single-family side informed how you approached this project?

ML: In mortgage, we think a lot about consumer experience and we are used to standardized processes. But our team had very diverse backgrounds — everything from engineering to open AI — and everybody’s expertise fit into the different pieces of what we were trying to pull together. So it was really cool to see how it all came together.

SW: What made you want to participate in this tech sprint, which was specifically focused on gen AI?

ML: Oh, I’m an uber nerd and I am super passionate about this! I host the TMC SparkLab show where we talk a lot about innovation in the industry. I do that because I’ve been in the industry for a little over 20 years and my parents were in it before that. I’ve had an opportunity to see over that span of time, how we as an industry have taken the pieces of paper that people have pushed across the desk to us and then created technology solutions that have facilitated that paper process. And the speed of that tech evolution in mortgage over the last 10 years has not kept up with the speed tech evolution everywhere else.

AI in general and specifically gen AI are starting to make inroads in the industry and help lenders to rethink their workflow design and what’s possible from a consumer experience and operational efficiency perspective. But all of these tools come with rewards and with extra risks. What I realized stepping into my role at TMC is that there really was an opportunity to help the industry see what both of those sides are, so that we can navigate this evolution the right way and protect our companies and protect consumers at the same time. That’s my “why.” So this was an opportunity to be a part of something that may be able to move the needle forward for the industry in some way, shape or form.

SW: What did you come away with from the tech sprint?

ML: There were a billion things! I learned a lot about multifamily. I learned a lot about open AI and chatGPT capabilities. I got the chance to learn from each one of my teammates as we pulled together the infrastructure for this use case.

SW: Does your use case require a human in the loop?

ML: No, that wasn’t built into our use case. It doesn’t seem as necessary in multifamily as it would be for the mortgage side of things — there’s a lot of complexities that happen in mortgage and a lot of risk that lenders look at with decisioning. So if a customer’s loan file can’t move through a very specific funnel, they need a human to be able to step in and help that customer. But it’s also so complex because somebody’s spending a lot of money over a long period of time and you need to make sure that consumers really understand what they’re doing.

In mortgage, we’ve been building very specific tracks that move things through the process, and because our industry is so complex, it’s hard for those tracks to account for every possibility. A self-employed borrower with 16 rental properties is very different than an FHA first-time homebuyer, so we’ve built specific tracks that can capture every scenario that consumers come up. But gen AI opens up the door for a different way of looking at things, because the system can learn and grow and digest information over much larger pools to be able to facilitate a transaction in a different way.

But, again, there are risks and rewards. There’s no magical gen AI solution that takes all humans out of the loop. On the multifamily side, they have all sorts of other risks that mortgage doesn’t have to think about — insurance for example.

SW: After your presentation, the judges gave notes and brought up the risk of fraud, specifically, fraud aided by gen AI. How did you guys think about that when you were developing the use case?

ML: One lens we used was the fraud that exists today. So, can we leverage a solution to help reduce that risk of fraud even by a fraction? Today, there’s some bias built into the tenant screening process because the tenant has to physically walk into an apartment building, but the fraud component is a little bit less if they’re having to physically produce documents. But our use case is really no different than in mortgage where a lot of the applications are online already. So we thought about it in that regard, too —  if we’re reducing the bias that a consumer would go through in the tenant screening process, that might help us to gain a little bit of incremental risk reduction.

SW: What was it like being part of this tech sprint? You guys were all doing this project on top of your regular roles!

ML: Yes, everybody was working at their regular jobs on top of working on our team. But we did have an opportunity to spend a lot of time together over a couple of days and everybody was really dedicated to making the solution something that we felt really great about. It was an opportunity to get to know folks really well when you’re put in that environment, which was awesome. Now we have a text group going with our team, where we’re still chatting about things. So that part’s fun and I loved the opportunity to network and meet new folks in the industry.

SW: The FHFA just started hosting a tech sprint last year. What do you think this kind of event can mean for innovation in the industry?

ML: This is an opportunity for us to explore possibilities. And anything we can do as an industry to reduce our cost of origination, reduce the risk involved in mortgage origination,  as well as selling it on the secondary market, that’s something I’m super passionate about exploring.

SW: I know your team was busy developing their own use case, but what did you think about the others that were presented?

ML: They were all interesting! I just happen to know that some of those things being presented are already being built, which means these were on the right track— they were looking at solutions that were really feasible in the marketplace.They came up with fantastic ideas that were real, tangible solutions for the industry. I have to be biased — I like ours the best, but there were 11 other fantastic solutions too.