Mortgage leaders discuss the future AI-powered workforce

September 15, 2025 6:00 AM

The rise of artificial intelligence has sparked both speculation and angst over what it might mean for mortgage professionals in the future, but the answers might be more encouraging than some expect.

Historically, advances in automation eventually brought transformations seen during the Industrial and Information Ages, which were marked by the elimination of certain positions. While the development of AI, machine learning and large-language models stand to make some jobs obsolete, mortgage technology leaders appear more likely today to view the technology's rise as an opportunity to optimize the current workforce and bring efficiencies into the workplace.

So far, research from various sources support the strategy. In new research conducted by Arizent, parent company of National Mortgage News, 50% of mortgage companies responding to a survey said talent optimization was their primary goal for technology adoption, while only 16% listed headcount reduction.

Similarly, in research looking across several industries conducted by the Federal Reserve Bank of New York, a growing share said they had ramped up AI adoption in 2025, but most did not plan to lay off staff as a result. A 47% share of service businesses expected to adjust to AI's breakthroughs by retraining staff, more than three times the 13% that expected it to cause downsizing, the central bank reported.

Several mortgage leaders spoke to National Mortgage News and expressed their views on the future look of an industry powered by AI. Many recognize this is a pivotal development that brings transformation but also the creation of new roles.

Responses were edited and condensed for clarity and length.

Ari Gross, chairman and chief innovation office, True

"You're going to have AI components, but who's conducting that symphony? I think you're going to have people in every area that are going to blend in the AI components making sure the whole system works for this company.

"There is orchestration required because you're going to have some bot, some human in the loop; it's going to be an amalgam of different things. AI is part of it, and who's going to conduct all of that?"

Rishi Choudhary, CEO, Kastle

"We are seeing two things. We've seen mortgage companies hire AI testers. Testing is still a big pain point. How do you effectively test AI agents? We're seeing mortgage companies either transition some of their supervisors into testers, and that's been really helping them assess the performance of these AI agents.

"The second is the prompt engineer — the builder who understands the business requirements and is able to use platforms like ours and build AI agents. That way, if a company is trying to transform their operations, they're able to do it fast."

Matt Rider, founder, Rider Consulting

Karen L Richard Photography

"I think what's interesting is how you are going to train your prompt engineers. There's no curriculum right now that I've seen that could evaluate what makes a good prompt engineer.

"That's where you've got to talk about data scientists too. I think when you get more into AI, you're going to see the need for more scientists and more model engineers. How are the large language models being constructed?

"I think there will be an entire organization to govern AI. Just like we have risk, just like we have information security and cyber, I think there will be an entire organization and a chief AI officer towards governing AI in these companies."

Chris Moschner, chief marketing officer, Finance of America

"I think this marketing technologist type role is actually emerging. It's a very interesting role where it's a person who, every day gets out of bed — they're not thinking about what to make — they're thinking about how to enable the team via technology and make sure all those things are working the way they should. They really act as a liaison with the tech teams.

"Product owners and tech owners are really worried about uptime, making sure the stuff works, making sure the data is right, making sure it's clean, whereas this technologist is really playing that intermediary role to make sure the team has that superpower that's enabled."

Randy Lightbody, head of mortgage, Truework

"Mortgages have been evolving over the last 50 years. AI is just the next step in this evolution; it's certainly not the last one. There's more coming beyond this. It's all about empowering the people to do what will help the firm be more successful; that is closing more loans faster, on time and with less exceptions.

"It's going to allow the current processors of the world to become — it could be exception manager, it could be something else. They'll have a new title, possibly, and that'll come with higher wages. They're going to make sure the AI does what the AI says it can do and be there to make sure the quality is top notch. It's empowering the people to use their expertise where it matters most."

Katherine Campbell, fractional chief marketing officer, Shape Software

"This is not necessarily a new role, but it's critical now that you have a strong technologist, and I mean a strong modern technologist. You need to have somebody that can come in and really understand all of these tools. Somebody can come in and say, 'the true recommendations are these tools, and this is how we're going to integrate them, and I'm going to have a regular relationship with our vendors so that I understand the roadmaps they're on for development.' That's new to mortgage.

"You have some companies that use 85 technologies just because they've been trying to appease LOs, and they come from different places. There's so much redundancy today because there has been so much money that has gone into these technologies for the capabilities that they have."

Michael Sachdev, CEO, Snapdocs

"What we are seeing in mortgage is that it's a regulated industry with a lot of structured data — some unstructured data as well — but it's a high velocity, high volume problem where accuracy really matters. What we see with LLMs is they do hallucinate, even with good prompts. If you've ever worked with a given LLM long enough, even if you've trained it, it actually will degrade over time.

"What I think that means for mortgage, at least absent some massive improvement in AI, is it's going to change roles. The way our customers are using our products is you have an employee who's doing a variety of tasks; some are more rote, and some are less rote. It is making it so they don't have to do the rote tasks. They're focused on exception management and more complex issues."

Steve Octaviano, chief technology officer, Blue Sage Solutions

"Too many people think of AI as the chatbot interface. That's how humans can use it. You can easily create a knowledge base and a chatbot that can answer questions, but everybody's doing that.

"But if you think about it, systems, applications, platforms are using AI. That's something that I don't hear people talking about as much, and that's important. The real value is when you connect it up to resources and tools within your platform, and then you start to use it to perform work automatically.

"AI is going to do the mundane. I think lenders that are smart will retain their staff and then use them for higher value positions, like advisory roles, customer relationship management roles and things of that nature."