What a real lender actually thinks about AI in mortgage

TL;DR: Mark Young, president and CEO of American Federal Mortgage, argues that most lenders misunderstand AI’s real value: they digitize manual tasks rather than eliminate them. This post covers how American Federal used AI workflow automation, including Ocrolus, to remove repetitive work from processors and underwriters, stabilize staffing through a cyclical market and improve borrower experience. Young predicts AI will follow the same adoption curve as eSign: a differentiator today, a baseline expectation in the near future.

When eSign technology arrived in mortgage, it was a competitive edge. Lenders who offered it won deals; borrowers who experienced it never went back. Within a few years, it became the floor. Any lender still routing paper was simply behind. Mark Young, president and CEO of American Federal Mortgage, believes AI is on the same trajectory. “For today I think it’s a differentiator,” he said, “but in the not so distant future it’s going to become something the borrower expects.” That window between early adoption and standard practice is closing. What lenders do in it will separate those who lead from those who catch up.

The hire/fire cycle in mortgage lending isn’t a strategy

The mortgage market is cyclical by nature. Rate movements, seasonal demand and broader economic shifts can compress or expand origination volume faster than most operations can absorb. The default response of scaling headcount up in busy periods and cutting when things slow is expensive, disruptive and hard to sustain.

American Federal Mortgage, a direct consumer lender based in New Jersey that closes 100 to 200 loans per month across the East Coast, takes a different approach. “We have never really been a hire/fire type company,” Young said. “It’s just not really the way we run a business and I don’t think it’s sustainable.” Instead of treating every market shift as a staffing problem, American Federal treats it as a workflow problem, which turns out to be a more solvable one.

With mortgage origination volume remaining sensitive to rate movements, lenders who have built flexibility into their operations are better positioned to absorb those swings without aggressive hiring or cuts. Workflow automation doesn’t eliminate market volatility, but it means the same team can handle more of it.

Digitizing processes isn’t the same as removing

This is where most lenders go wrong. Digitizing a workflow means giving a manual process a digital interface: moving a checklist from paper to a portal, routing a document through email instead of fax. The task still exists. The time it consumes doesn’t change.

Removing a task is something different. “We saw the difference in AI applied to mortgage because it was actually removing manual tasks from our process rather than just digitizing a workflow,” Young said. “We really were removing repetitive tasks.”

That distinction matters because digitized tasks still consume processor and underwriter time, still create bottlenecks and still slow the borrower down. American Federal deployed AI tools, including Ocrolus, to eliminate document-handling work that had previously pulled processors away from borrower-facing activity. Once those tasks were gone, not just moved into a new system, the team’s capacity changed in ways borrowers could actually feel.

When processors stop doing repetitive work, borrowers notice

American Federal’s model is built around a single point of contact for every borrower: one representative from application through close. It’s a meaningful differentiator, but it only works if that rep is available and focused, not buried in back-office paperwork.

Removing repetitive tasks made that possible. “Our processors can focus more on customer service and speaking to the borrower,” Young said. “That’s resulted in a better borrower experience and hopefully down the line, turn to more referrals and repeat business for us.”

Young also noted something worth highlighting about change management: when results are real and visible, adoption builds on itself. “Your team gets excited about it and then they’re looking for different things to either automate or use tools to make more efficient.” The flip side holds equally: partial adoption undermines the whole exercise. “If you’re adopting these workflows and you’re still doing the manual process at the end, that doesn’t help anybody and it actually becomes less efficient.” Leadership has to commit to the full transition, not just stand up a tool.

The lenders who make that commitment now are building a lead their competitors will struggle to close. Borrowers already expect faster answers and fewer repetitive documentation requests, and that expectation will only sharpen. The tools that differentiate today have a way of becoming the baseline expectations of tomorrow.

Key takeaways

• AI in mortgage is following the same adoption curve as eSign: what sets lenders apart today will become what borrowers expect in the near future.
• The cyclical mortgage market makes headcount-driven scaling unsustainable; AI workflow automation offers a more durable way to absorb volume changes.
• Digitizing a manual process and removing it are fundamentally different; only elimination actually frees up processor and underwriter capacity.
• When repetitive tasks are truly gone, the direct beneficiary is borrower experience; processors can focus on relationships rather than documents.
• AI adoption fails when leadership doesn’t fully commit; teams executing the manual fallback at the end of an automated workflow lose the efficiency gain entirely.

FAQs

What is the difference between digitizing and automating a mortgage workflow?

Digitizing means moving a manual task into a digital environment: a paper form becomes an online form, a fax becomes an email attachment. The task still exists and still requires human time. Automating, in the meaningful sense, means eliminating the task entirely or removing the human step from it. For mortgage lenders, the difference shows up in whether processors are freed to focus on borrowers or still spending time on document handling that moved to a new interface.

How does AI help mortgage lenders manage volume volatility without adding staff?

AI workflow automation allows a fixed team to process more volume by removing the repetitive, time-consuming tasks that create bottlenecks during busy periods. Rather than hiring into peaks and cutting in slow periods, a cycle that’s expensive and disruptive, lenders can build flexibility into their workflows. When volume spikes, the same team can absorb it without adding headcount.

How does removing manual tasks improve the borrower experience in mortgage?

When processors and underwriters aren’t tied up with repetitive document handling, they have more time and attention for borrower communication. For lenders like American Federal Mortgage, that means the single point of contact model, one representative from application through close, actually works in practice. Borrowers experience faster responses, fewer redundant requests and more consistent service.

When will AI in mortgage shift from competitive advantage to standard expectation?

It’s already happening. Mark Young of American Federal Mortgage draws the parallel to eSign: once a differentiator, now a baseline expectation. His view is that AI-powered workflows that deliver immediate results for borrowers will follow the same path. Lenders who wait for AI to become standard before adopting will be catching up, not competing.

What does successful AI adoption require from mortgage leadership?

Commitment to a full transition, not a partial one. Deploying an AI tool while keeping manual fallback processes in place doesn’t reduce workload; it adds to it. Successful adoption requires leadership to set the direction, give the team time to adjust and stay the course long enough for results to become visible. Once team members see genuine improvements in their daily workflow, adoption tends to accelerate on its own.