Artificial intelligence (AI) is no longer some far-off futuristic concept—it is already changing how we work in mortgage servicing. From an industry perspective, we are focused on the same goals: staying compliant, keeping costs under control, serving customers well, and holding onto those customer relationships over time. What has changed is that we now have a new set of tools to help us hit those goals, and AI is proving to be one of the most powerful tools we have ever had.
I call AI a “mega-trend” because it is not just another technology upgrade, it is a shift that is impacting nearly every industry and every corner of society. In servicing, the impact is clear: lower costs, better risk management, faster processes, and more personalized customer experiences. And we are still just scratching the surface.
Three Types of AI
To understand AI’s role in servicing, it is helpful to distinguish between three categories of capabilities: Predictive AI, Generative AI, and Agentic AI.
Predictive AI is the most mature and widely used form of AI. By analyzing large amounts of historical and real-time data, predictive models can forecast outcomes and highlight risks. In servicing, predictive AI is helping with analytics, compliance checks, loan boarding, and customer retention. For instance, it can look at customer payment patterns and help servicers identify which loans are at risk of delinquency so they can proactively reach out to customers to offer assistance. This represents a shift from reacting after a problem occurs to anticipating and addressing issues before they escalate.
Generative AI is still in its early stages, but it is quickly becoming valuable. It can draft reports, summarize regulations, and even produce customer communications. Tasks that once took days or weeks—such as translating new FHA or GSE guidelines into operational playbooks—can now be accomplished in a fraction of the time. Generative tools are also being used to create investor updates or compliance summaries, saving staff countless hours.
Then there is Agentic AI, which I believe represents the next evolution. Instead of just automating tasks programmed in advance, agentic AI can make simple decisions on its own and push those decisions back into servicing workflows. That is a big deal because it moves AI from being a passive tool to an active participant in day-to-day operations.
Smarter Conversations
The most mature applications of AI in servicing so far have been internal—loan boarding, document ingestion, and compliance controls. In these areas, AI has already sped up processes and improved accuracy.
Customer communication is a more recent development in the use of AI. Chatbots and virtual assistants are now answering frequent questions like “What’s my escrow balance?” or “When’s my payment due?” at any hour of the day. Customers do not have to wait on hold, and call centers are not bogged down by repetitive questions.
Over the past few months, in fact, we have launched a pilot program in which AI agents interact directly with customers through both chat and voice. The system can handle up to 100 conversations at once, which means no one must wait in line behind another customer. It is not just about scale, but effectiveness. We can measure whether a customer who used our system calls us back within three days, and between 40% and 70% of the time, they do not. This tells us the system is doing its job.
Though the program is relatively new, the early results are promising. Since we introduced this offering, we’re seeing a 73% resolution rate of customer requests, and 400 fewer calls per week into our call center. Customers are engaging well, too. For servicers, that combination—better customer experience and lower cost—is powerful. It has worked so well that we are expanding the pilot.
It is important to note that this technology is not about replacing people. Instead, it makes human staff more productive. By handling simple and repetitive questions, AI frees up employees to focus on the more complex cases that require human judgment and expertise. It’s similar to moving from a typewriter to a computer—the work doesn’t disappear; it just gets done more efficiently.
Customer Retention and Other AI Applications
When I think about other exciting uses for AI in servicing, I immediately think about customer retention. Retaining a customer is about more than just keeping an account on the books—it is about maintaining trust and being top-of-mind over the life of the loan. As servicers, we are not only managing day-to-day obligations like escrow or payment processing, but we are also competing for relationships in an environment where customers have options.
Over the past year, we have been using AI to get better at this. Technology helps us answer two critical questions: which customers should we be talking to, and what should we be saying? Traditionally, retention efforts relied on broad campaigns—generic letters or emails that went out to everyone. With AI, we can move away from that one-size-fits-all model. Instead, we can identify which customers are most likely to refinance or pay off their loans early and tailor our outreach accordingly.
Our prepayment score predictor model plays a key role here. By generating a score for each customer based on factors like market rates, loan age, and homeownership history, we can pinpoint which customers are most at risk of leaving. Instead of chasing the entire portfolio, we can focus resources on the segment that may be considering refinancing or moving. That not only saves money but also enables us to have more meaningful conversations with customers who are actively weighing their options.
Of course, AI does not stop with the scoring. It supports the messaging itself. Whether it is helping refine copy for customer communications or assisting our marketing team in testing different approaches, AI allows us to continually improve our outreach. We are not just sending out a message and hoping it sticks; we are learning from each interaction and adjusting in real-time.
For me, this is one of the most practical, mature applications of AI in servicing today. It does not require futuristic breakthroughs. It simply takes data we already have, applies intelligence to it, and gives us insights we can act on. The result is stronger customer relationships, fewer surprises, and a more resilient portfolio. And in a market where churn can be costly, that is a very real advantage.
Default management and loss mitigation are also being transformed. In addition to determining whether a customer might be at risk, AI also makes document-heavy processes much faster. AI-based tools can pull the right data from hardship letters, income statements, or workout applications in seconds. Decision-support systems can then suggest which solution—modification, forbearance, or repayment plan—makes the most sense for both the customer and the investor.
On the operational side, AI is cleaning up a lot of the manual work. It can spot errors in tax bills or insurance renewals before they become compliance problems. Predictive models can also flag which transactions are most likely to fail, so teams can focus their time where it is really needed. Meanwhile, Generative AI can draft investor reports or compliance updates quickly and clearly, eliminating hours of manual writing and formatting work.
AI’s Two Essentials
The benefits of AI are already clear. Loan boarding that used to take three, five, or even seven days can now be completed in just one day with full compliance. Customer retention efforts are more effective, since AI can pinpoint the 20%-30% of a portfolio that is responsible for most payoffs. And with AI chat and voice agents, customers get their questions answered more quickly, and employees can spend more time on the most important work.
However, technology alone is not enough. Two other factors are critical for success with AI: data and people.
Data quality is the foundation for any effective AI solution, since AI models are only as good as the information being used to train them. The human factor is just as important. Because employees will naturally ask whether AI will replace them, leaders need to make sure they understand that AI is here to help make their work easier. That means being able to explain what is being done and why, showing teams how they will benefit, and answering their questions.
For instance, at our company, we have expanded access to AI tools to employees at all levels, which has created a sort of openness that helps build trust and speeds adoption.
Where Are We Headed?
The biggest thing to understand about AI is that it gets better the more you use it. Every customer interaction, every compliance check, and every document processed makes the system smarter. That means early adopters have a chance to build an advantage that compounds over time. On the other hand, organizations that hold off on implementing AI may find themselves playing catch-up because their systems will not have the same experience or learning built in.
The question is no longer whether AI will be used in servicing—the question is how well it is used. The servicers who succeed in the coming years will be those who use AI to strengthen compliance, deepen customer relationships, and empower their teams. The best of them will treat AI as a partner, invest in clean data, and bring their people along on the journey.
AI is not an end in itself—it is a tool. But it is a tool unlike any we have had before, one with the power to redefine servicing and reshape customer expectations.
