Voice of the Industry: David Spector (Part 6)

Over the past couple years, I’ve had the opportunity to sit down with Pennymac Chairman and CEO David Spector on multiple occasions, and each conversation has offered a deeper look into both the company and the broader mortgage industry. Our most recent conversation was shortly after the 5th Annual Stanford L. Kurland Memorial Golf Classic, where the industry once again came together to raise millions of dollars for brain cancer research, an event that has become one of the mortgage industry’s signature philanthropic efforts. This time, our discussion focused on the future of mortgage servicing. With Pennymac’s pending acquisition of Cenlar, accelerating advances in artificial intelligence, and continued consolidation across the servicing landscape, Spector shared why he believes controlling technology is increasingly becoming a competitive advantage, how AI can improve both customer experience and regulatory compliance, and where he sees the servicing industry heading in the years ahead. What follows is our edited conversation.

Q: I recently saw you at the Stan L. Kurland Memorial Golf Classic, where the industry raised a tremendous amount for brain cancer research. What does that event mean to you?

A: Over the past five years we’ve raised $13 million to help fight brain cancer, and I’m incredibly proud of what the UCLA Neuro-Oncology Program is doing as a leader in this area of research. I’m also proud of how our organization has rallied behind this cause.

What impresses me most is the continued generosity of the people who attend. Every year we raise more money and find ways to refresh the event. Philanthropy is something this industry does very well, and like many large mortgage companies, we’ve found a cause that’s deeply meaningful to us.

Q: Looking at today’s mortgage servicing landscape, what are the biggest challenges facing the industry?

A: There are a few. First, industry consolidation has resulted in servicing portfolios being owned by larger, better-capitalized firms, which is generally positive. My concern is that not everyone has invested in servicing the way we have. Some firms view servicing simply as something they have to do rather than as a business worth investing in.

A good example is the VA Partial Claim program. We will be the first large servicer ready to offer it, while some competitors are still months away. We were able to move quickly because we own servicing assets ourselves and directly benefit from investing in the platform.

The second challenge is technology. Over the last five years, many servicing technology providers have focused more on reducing fees than modernizing their systems. When the industry’s primary goal becomes lowering costs, technology investment inevitably suffers. You end up with platforms that struggle to support increasingly complex loan types. We’ve taken a different approach by investing heavily in technology, while still remaining competitive on price, allowing us to offer capabilities we believe no one else in the industry currently provides.

Q: Servicing technology has to support borrowers throughout the life of a loan. Why is that so important?

A: It starts with the customer. My first, second, and third priority is always the homeowner’s experience.

Consumers need multiple ways to interact with their servicer, whether they’re checking a payment, asking about taxes or insurance, dealing with job loss, divorce, the passing of a loved one, or requesting loan modifications. The technology has to accommodate all of those situations while respecting people’s time. Consumer expectations today are shaped by every digital experience they have. If a website requires too many steps to accomplish something simple, people leave.

Different customers also have different preferences. Some want to speak with a person; others would rather interact through a natural-language virtual assistant. Someone who’s going through a hardship might request to handle these types of things digitally rather than discussing it with another person.

Behind the scenes, the platform also has to satisfy investors, regulators, internal audit, quality control, and numerous operational requirements. That demands sophisticated workflows. Those workflows are where the real efficiency gains come from.

We’re the only major servicer reducing our cost to service, while many competitors are seeing costs rise. Others often respond to increasing complexity by hiring more people or bolting on additional software modules. We believe the better solution is building integrated workflows that automate the process from the beginning.

Q: Servicing is obviously a huge part of Pennymac’s business. How is the acquisition of Cenlar progressing?

A: It’s going very well. We expect to close sometime in the fourth quarter, and the teams are coordinating exceptionally. Our primary focus is ensuring that Cenlar’s roughly 100 clients experience as little disruption as possible – that’s what we mean by “zero lift.” As clients learn more about the transaction, they’re recognizing the value of having Pennymac behind the platform. They’ll benefit from our technology, our stronger balance sheet, and a team that actually originates mortgages, owns servicing assets, and invests its own capital alongside clients. That creates much better alignment.

This isn’t a traditional M&A transaction where you realize immediate cost synergies on day one. The efficiencies will come as clients gradually migrate from Cenlar’s technology onto our platform. Overall, the integration is progressing very well, and I’m excited to get the deal closed and move forward.

Q: Why does Pennymac believe in owning its technology, Plaisse, rather than relying on third-party systems?

A: Our philosophy is simple: we want to control our own destiny.

On the production side, we do use a third-party platform, but we invested in that provider because we wanted a seat at the table. That gave us confidence that we could influence the product’s direction rather than simply relying on someone else’s roadmap. We’ve successfully deployed that technology in our consumer-direct channel, we’re leveraging it for AI initiatives, and now we’re rolling it out to our third-party origination channel.

Servicing was different. We concluded that third-party servicing platforms weren’t keeping pace with the industry’s needs or investing sufficiently in innovation. At the same time, we found ourselves building more and more technology around those systems. Eventually we decided it made more sense to finish the work ourselves and own the core platform. Plaisse was built by servicers, for servicers drawing on Pennymac’s experience managing complex mortgage operations at scale. It’s not about perfection; it’s about controlling as much of your future as possible to provide homeowners the best experience.

Q: How does subservicing fit into Pennymac’s long-term strategy?

A: We’ve actually been subservicing since 2009 through our REIT, PMT, although that was an affiliated relationship. We expanded into third-party subservicing about four years ago because we understood the business from every angle.

We know what servicing investors need because we’re servicing investors ourselves. We also understand what the agencies and other investors require in terms of reporting, remittances, advances, and compliance.

Growing the subservicing business allows us to offer our technology to others in the industry, increase scale on our platform, generate additional resources to reinvest in technology, and earn a return on the substantial capital we’ve invested in building the system.

When the industry improves, everyone benefits.

Q: Where do you see automation and artificial intelligence having the biggest impact in servicing?

A: The biggest impact is behind the scenes.

Automation is extremely valuable for functions like bank reconciliations and providing real-time assistance to customer service representatives during calls through live transcripts and suggested responses. AI is also transforming quality control by allowing us to analyze enormous amounts of servicing data using large language models, improving both consistency and regulatory compliance.

Historically, when new work emerged in mortgage servicing, the answer was often, “Just hire another person.” That’s expensive, but more importantly, it’s not a sustainable long-term solution. Every additional human introduces another potential point of inconsistency.

What we’re seeing with AI is much greater consistency in customer service. Homeowners receive the same quality of information regardless of which representative they interact with. Because everything is measurable and data-driven, we actually expect AI, when implemented correctly, to improve compliance rather than weaken it.

That’s another reason I’m excited about bringing these capabilities to Cenlar’s clients; it should help create a more disciplined, compliant servicing environment.

Q: What should mortgage companies consider before implementing AI in servicing?

A: First, don’t be afraid of it.

I tell our teams to treat an AI agent in the same way you’d delegate work to someone on your team. You can hand off tasks, but must be responsible for the outcome. You can delegate authority, but you can’t delegate responsibility.

The second point is that technology teams shouldn’t be the only people building AI agents. Business users should have platforms that allow them to create, test, and deploy AI themselves. The old model – where business teams wrote specifications, handed them to technology, waited for development, tested, sent changes back, and repeated the cycle – is over.

If the platform is designed correctly, the business should be able to build and deploy AI directly. That’s incredibly powerful, but it also makes business leaders fully accountable for the results. In many ways, it’s no different than building a spreadsheet with macros, except today’s AI agents are vastly more capable.

Q: There’s increasing discussion about designing technology for the “happy path”—straightforward loans with no complications. How should servicing technology handle more complex situations like loss mitigation?

A: Building a servicing system for a borrower who always makes payments on time is easy.

The hard part is building a platform that supports every borrower through every possible circumstance. That’s what servicing is really about.

I’ve heard large banks say they love servicing current loans but aren’t very good at servicing defaulted loans. Well, processing an automatic payment isn’t difficult. The real challenge comes when borrowers face financial hardship, make late payments, dispute taxes or insurance, or need assistance.

If you’re going to be in servicing, you can’t just focus on the happy path. You have to be all in.

Q: Looking ahead, where is Pennymac—and the servicing industry more broadly—headed?

A: We intend to continue growing our servicing portfolio while expanding across every part of the servicing ecosystem. That includes owning servicing assets, servicing our own loans, providing subservicing to other institutions, and helping clients with additional servicing solutions.

We want to leverage our experience and what we believe is a best-in-class management team to improve the industry as a whole.

If interest rates remain elevated, I expect another wave of consolidation. The industry will likely continue concentrating among larger, well-capitalized institutions. I also wouldn’t be surprised to see banks gradually re-enter mortgage servicing, particularly on the conventional side, though perhaps not government servicing initially. At the same time, smaller independent mortgage banks will likely continue to see their influence diminish.

Q: On a lighter note, when you’re driving, are you listening to music, podcasts, or taking calls?

A: I have four kids, so they know they can usually reach me between 8:00 and 9:00 in the morning and again between 4:00 and 5:00 in the afternoon, although that afternoon window depends on how late I’m working. During the mornings I’m usually on work calls, and after that I like listening to podcasts.

Q: Any favorite podcasts?

A: If I can get through the Chrisman Commentary, I consider it a success.