Access the Messagepoint Video

Pressures Facing Mortgage Servicers

As of Fall 2024, Mortgage servicers are in a tough spot. As Duwaine Thomas put it bluntly, “Household debt has spiked, interest rates remain high, and home affordability is worse than ever. That puts real pressure on the customer—and on us.”

Mortgage delinquencies are on the rise, with the cost of servicing non-performing loans estimated to be 11 times higher than that of performing ones. This spike in borrower distress has forced servicers to increase the volume and complexity of outbound communications—letters, disclosures, notices—many of which must be highly customized and compliant across states.

Making matters worse, servicers are operating with legacy tools and fragmented processes that hinder speed, drive up costs, and increase compliance risk. “You’ve got letters being built in Word, clunky legacy platforms, and teams sending content back and forth with print vendors—it’s slow, expensive, and not sustainable,” Kehoe noted.

A Primer on Using AI in Mortgage Servicing

The webinar then shifts to explain what artificial intelligence actually is, and how servicers can best leverage it as part of their borrower communications process. Kehoe offered a grounded explanation: “At its core, generative AI is just a sophisticated math model that predicts the next most likely word given a prompt. But the power lies in how you shape that prompt and integrate it into your workflow.”

This distinction matters. Using ChatGPT in isolation might create more risk than value. But when embedded directly within content systems and governed by enterprise-grade safeguards, AI becomes a force multiplier—improving quality, accelerating change cycles, and reducing costs. “AI isn’t about replacing people,” Kehoe emphasized. “It’s about giving them superpowers.”

3 ways AI can help reduce borrower communications costs

That conceptual foundation sets the stage for a more tactical discussion: where exactly can AI make the biggest dent in servicing costs? The panel identified three high-impact areas

Reducing Calls to the Call Center

A major cost driver for servicers is avoidable call volume—especially when unclear or poorly timed communications send confused borrowers straight to the phone. As Kehoe pointed out, “If your content is hard to understand, people call. And that’s expensive.” According to Duwaine Thomas, call center, collections, and loss mitigation costs can account for up to 45% of the cost per loan—and climb even higher when defaults surge.

One root cause is complexity. Most borrower letters are written at a grade 13 or 14 reading level, while half the U.S. population reads at grade 5 or below. Add in regulatory language, inconsistent templates, and translation gaps, and it’s no surprise customers can’t self-serve.

Generative AI can help close this gap by simplifying content. AI can analyze documents for readability and automatically rewrite them to meet internal standards or ISO-defined plain language guidelines. “You’re not just getting a dumb rewrite,” said Kehoe. “You’re applying rules—grade level targets, sentence structure, clarity—to ensure consistency and control.”

AI can also dramatically improve translation workflows. Instead of translating only a handful of documents due to cost and complexity, servicers can now translate entire libraries—accurately and in near real time. AI doesn’t just translate; it validates content alignment, checking for consistency in variables, tone, and intent across languages. “It’s not just faster,” Thomas added. “It’s better. We’ve seen AI catch errors that got past certified translators.”

Loss Mitigation

Nowhere is the demand for speed and clarity higher than in loss mitigation. Borrowers in distress need to understand their options—fast. But change cycles are slow and risky. “We’ve seen cases where changing a phone number took weeks—or even months,” said Thomas. “And in that time, incorrect letters went out, customers couldn’t get through, and complaints poured in.”

AI and modular content help break this cycle. Rather than hard-coding text into dozens of templates, modern systems store reusable content components—disclosures, state variations, instructions—in a centralized library. When a change is made, it cascades across all touchpoints instantly.

For example, a change to a single loss mitigation clause could trigger 26 downstream updates in traditional systems. With a centralized content hub that enables content to be shared between communications, it’s one edit, applied everywhere. “This is how you go from reactive to agile,” Kehoe said. “You’re not waiting for IT, not locked into vendor timelines. You’re in control.”

Servicers can also tailor content by state, investor, or brand—all without creating document sprawl. The system applies rules to dynamically assemble the right version for each borrower.

Reducing Print and Mail Costs

Despite consumer preferences shifting digital, mortgage servicing remains heavily print-reliant. JD Power data shows that 43% of borrowers receiving printed statements actually review them online—a clear opportunity for cost reduction.

The barrier to going digital is technological for most servicers. “Most digital communications today are PDFs,” said Kehoe. “They’re impossible to read on a phone. So borrowers don’t opt in.”

AI can help accelerate digital transformation. By automatically reformatting content for email, SMS, or mobile app delivery, servicers can offer better experiences without rebuilding each message manually. “It’s the same core message, adapted to the medium,” he explained. “And that shift can save you six figures a year in postage alone.”

Next Steps

To realize these benefits, the panel advised starting with the content itself. Centralize it. Modularize it. Remove the barriers to reuse and automation. Then, apply AI in specific, high-impact areas like readability improvement, translation, and multichannel formatting.

“You don’t need to overhaul your whole stack overnight,” said Kehoe. “Start with borrower communications. The ROI is tangible, and the operational gains are immediate.”

Thomas echoed that sentiment: “This isn’t about tech for tech’s sake. It’s about getting better outcomes—fewer calls, faster turnaround, clearer options for the borrower, and real savings.”

Whether you’re just beginning to explore AI or already using it in limited ways, the message was clear: Start small, but start now. The pressure on servicers isn’t going away and those who are able to successfully integrate these tools into their borrower communications processes will gain a competitive edge moving forward.

This is just a preview of the insights shared in this webinar. Complete the form above to watch the full recording.