It’s difficult to understate the impact that artificial intelligence is having on all aspects of the mortgage industry. From lead generation and borrower qualification to loan servicing and customer retention, companies must find ways to streamline processes to bring more customers in the door and keep their bottom line healthy.

At HousingWire’s The Gathering in Austin this week, AI was a central theme. Here are some of the ways executives and front-line personnel can look to effectively deploy AI tools on an internal or external basis.

Pennymac debuts rebranded servicing tech

David Spector, chairman and CEO of California-based PennyMac Financial Services, on Wednesday introduced the audience to the company’s proprietary, cloud-based AI servicing technology. The system has been rebranded as “Plaisse,” a French phrase that means “to please.”

“Plaisse was developed ‘by servicers, for servicers’ to address the unique challenges of our industry with the goal of enabling efficiencies in the mortgage loan servicing journey with a best-in-class experience for homeowners,” Pennymac explained in a social media post.

The top five U.S. mortgage lender is also growing its AI presence in the originations space. In October 2025, Pennymac announced its purchase of an equity stake in Vesta and became the first major lender to join the Vesta loan origination system platform.

Spector said the move was designed to give Pennymac a legacy-free, state-of-the-art LOS run by experts. The partnership was fully deployed as of February, he said, and is already producing benefits through faster speeds and lower costs to close.

While Vesta and Plaisse aren’t merging, having two modern AI-based platforms will enable better internal communication and customer service. “The production and servicing worlds are just moving closer and closer together. These worlds are colliding,” Spector said.

Fairway’s massive AI deployment

Leaders at Fairway Home Mortgage discussed their company’s two-year enterprise AI journey with TRUE, a loan decisioning software provider. TRUE CEO Steve Butler was the moderator for the session.

Fairway’s Caleb Ondrusek and Deedra Massey Bosworth said their company made a deliberate decision about three years ago to build a strong data foundation, rather than focusing on AI solutions for mid-process or post-closing tasks.

They told the audience that AI has great potential to transform the mortgage process, but doing it successfully hinges on having an intuitive user experience and a willingness to take measured risks.

“We kind of realized we have to start at the beginning … make sure that we have solid data at the beginning that matches all the documentation that the borrower provides, and then we can carry that forward throughout the process,” said Bosworth, Fairway’s senior vice president of technical product management.

Fairway leaders said they went all in on enterprise AI solutions rather than isolated pilot-use cases. They’ve deployed these across more than 600 branch locations and are being used by more than 2,000 loan officers. The larger scale reveals key dynamics that smaller deployments cannot, they said.

“That enterprise level really allows us to develop a process,” Bosworth said. “Our idea is to eliminate steps in the process, not add steps.”

Ondrusek, the company’s executive vice president of technology and innovation, said that change management is at the heart of a successful AI deployment since the industry generally has a poor track record of technology adoption. The message from lending executives should be to set intent while affirming they want to augment human tasks rather than replacing people.

“If I had some advice for people just starting out on this journey, it’s really going to come down to having a workforce plan before you start,” Ondrusek said.

CreditXpert: Can LOs compete with chatbots?

Mike Darne, vice president of marketing at CreditXpert, led a session where he argued that AI-powered chatbots like ChatGPT and Claude have become de facto agents for prospective borrowers.

These tools can often coach consumers on credit scores, loan-level price adjustments, mortgage rate impacts and the mechanics of a transaction before they ever speak to a loan officer. Darne said this is not a passing fad but a structural shift in how consumers enter the mortgage funnel. Lenders should respond with better execution, not just better marketing.

Darne said LOs should meet clients with a holistic financial plan rather than a prequalification letter, which he said often serves as “an invitation to go out and shop.”

“This guy I talked to last week, he said, ‘Any time I get a lead, my assumption is that this person is aggressively shopping out there. And I’ve got to respond. I’ve got to build trust,’” Darne said.

He went on to explain that credit repair is a human-driven process that often achieves lower-quality results as more borrowers leave the sales funnel and don’t return.

Conversely, he argued that credit optimization can help 60% to 70% of borrowers achieve a better score within 30 days, per CreditXpert data. The company uses predictive analytics and machine learning built with 25 years of data to generate precise improvement plans and communicate the likely outcome to the borrower upfront.

“We’re not talking about a niche part of the market,” Darne said. “Over the past couple months, I’ve been spending a lot of time just going out and speaking to our clients at the end-user level. And the ones that are winning are doing just this.”