Surviving the Next Refi Boom with Matt Chojnacki

This is a summary of an episode of Pioneers, an educational podcast on AI led by our founder. Join 3,000+ business leaders and AI enthusiasts and be the first to know when new episodes go live. Subscribe to our newsletter here.

TL;DR:

• Lenders that adopt agentic AI now will be better equipped to handle the next refi wave without hiring massive ops teams.
• Agentic AI can automate underwriting decisions, triage documents, and surface red flags, freeing teams to focus on edge cases.
• Mortgage tech stacks are fragmented; lenders want AI that integrates into their LOS and workflows without replacing core systems.
• Automation improves not only efficiency but also makes lenders more attractive to top loan officers and underwriters.
• Waiting for volumes to rise before adopting AI is risky. AI readiness is a long-term advantage, not a last-minute fix.

Before we dive into the key takeaways from this episode, be sure to catch the full episode here:

Meet Matt - Account Executive at Multimodal

Matt Chojnacki, Account Executive at Multimodal, brings deep experience in lending and mortgage operations to the AI space. He started his career as a loan officer at Quicken Loans during the peak of the last refinance boom and later pivoted into mortgage-specific CRM and tech sales.

At Multimodal, Matt applies that front-line experience to help lenders future-proof their operations using agentic AI. By engaging daily with lenders, he offers insights into the real challenges facing loan officers and underwriting teams—from document overload to hiring bottlenecks.

An advocate for AI that actually works in production, Matt partners with enterprise teams to show how agentic AI can automate routine tasks, surface red flags, and keep humans in the loop, making scaling possible without sacrificing compliance or trust.

Why the Next Refi Boom Will Be Won With AI, Not Headcount

Matt explains that during the 2020–2021 refi boom, lenders scaled by hiring aggressively. But that approach isn’t sustainable.

When interest rates dropped, mortgage companies staffed up fast, only to lay off thousands when volumes dried up. Now, as rates start to fall again, lenders are exploring smarter ways to scale without repeating the past. AI Agents offer a better path.

They can handle repetitive tasks, reduce bottlenecks, and allow humans to focus on complex cases.

This shift enables lenders to stay lean while still meeting market demand. Instead of scrambling to build massive teams, the lenders that deploy AI today will be able to move quickly, capture volume, and avoid the burnout and chaos of past cycles.

From Ops Bottlenecks to AI Agents: What’s Changed Since 2021

Back in 2021, lenders were overwhelmed by application volume.

Matt recalls how the operations teams were flooded, documents piled up, and turn times slipped. But since then, the tech landscape has evolved.

Lenders are no longer just looking for automation. They're looking for agentic AI: systems that understand context, perform multistep tasks, and triage intelligently.

AI Agents are not just task bots. They operate like teammates, freeing staff from repetitive work and reducing the need for last-minute staffing spikes.

As Matt explains, the pain of 2021 pushed the industry to rethink how lending should scale. Today’s AI tools are more mature, safer, and easier to integrate. Also, lenders are now ready to use them as part of their core operating model.

Automating Underwriting: How Agentic AI Helps Without Replacing Humans

Matt is clear: agentic AI isn’t here to replace, but support underwriters. In the lending space, automation often hits resistance because people assume it means job loss.

But Matt sees something different. Underwriters today are stretched thin, managing complex borrower scenarios and high expectations.

AI Agents can take care of the first pass—pulling documents, highlighting risks, identifying missing information—so that underwriters spend more time on edge cases and less time on routine checks.

Matt compares it to giving them a digital junior teammate. This not only improves turn times but also reduces burnout. Instead of eliminating roles, agentic AI lets underwriters focus on judgment work where their expertise really shines.

Document Overload? How AI Can Triage, Flag, and Recommend

In a typical loan file, hundreds of documents might flow in, and Matt says that loan teams are drowning in the volume. Traditionally, staff have to open, read, sort, and validate each file.

That creates delays, inconsistencies, and risk. AI Agents can take over this process by reading documents as they come in, classifying them correctly, and flagging issues before humans even see them.

“Document overload slows everyone down—AI helps triage and flag in real time.” — Matt Chojnacki

If an income doc is blurry or a W-2 is missing, the AI flags it, notifies the team, and even recommends follow-up actions.

“Agentic AI is about action, not just answers.” — Ankur Patel

This kind of triage helps underwriters and processors focus on the work that actually requires human review. It also improves speed and consistency across the board and reduces the risk of costly oversights.

Why the Smartest Lenders Are Preparing Now, Not Later

Matt emphasizes that lenders who wait for volume to spike before adopting AI are making a strategic mistake.

Building AI readiness takes time—time to test workflows, ensure compliance, and get internal teams up to speed. If lenders delay until the market heats up, it’ll be too late to scale effectively.

Those who are experimenting now are better positioned to capture future volume without burning out their teams. Matt compares it to preparing for a game: you don’t wait until kickoff to start training.

Enterprises that take steps now to embed AI into their operations will avoid the staffing scrambles and tech debt that held them back during the last cycle.

Beyond RPA: Why Agentic AI Is a Step-Change for Mortgage Tech

Matt distinguishes clearly between yesterday’s RPA tools and today’s agentic AI. RPA was great at moving data from one place to another or clicking buttons in bulk, but it didn’t understand context.

It couldn’t read documents, surface exceptions, or collaborate with a human team. Agentic AI, by contrast, can interpret inputs, make suggestions, and manage workflows proactively.

“Agentic AI isn’t here to replace people. It’s here to support them.” — Matt Chojnacki

It understands not just what to do, but why it’s being done. Matt explains that lenders don’t want another standalone tool. They want AI that can reason through underwriting scenarios, support compliance, and plug into their daily systems.

This is a fundamental shift from automation to augmentation—and it’s why agentic AI is winning adoption where legacy automation stalled.

Hiring? AI Makes You a Magnet for Top Loan Officers

Matt points out that AI doesn’t just improve operations—it also helps attract and retain talent.

Loan officers want to work for companies that support them with modern tools and efficient systems.

If every deal takes days to process because ops is backed up, high-performers will leave. But when AI Agents handle admin work—collecting documents, validating data, triggering next steps—loan officers can close faster and focus on relationships.

That efficiency boosts morale and earnings. Matt says that today’s most competitive lenders use tech as a differentiator in recruiting. Being known as an “AI-forward” lender signals to top LOs that they won’t be stuck fighting bottlenecks all day. Instead, they’ll be empowered to do their best work.

How AI Agents Integrate Without Replacing Your LOS

One of the biggest barriers to AI adoption in lending is the fear of replacing core systems like the LOS (Loan Origination System). But Matt clarifies that agentic AI doesn’t require ripping and replacing existing infrastructure.

Instead, these agents sit alongside your existing stack, integrating with LOS platforms, CRM tools, and internal workflows.

“You don’t need to rip out your LOS. AI fits right in.” — Matt Chojnacki

Matt emphasizes that lenders want augmentation, not disruption. AI Agents can trigger tasks inside your LOS, pull data, send updates, and manage exceptions, all without requiring massive IT projects. That low-friction integration is why adoption is growing.

Lenders can get the benefits of automation without the pain of replatforming, making it easier to test, iterate, and scale AI in production.

Want to learn more about AI-driven lending? Check out this episode featuring 10 key takeaways on AI-driven lending with Andy Mattingly.