The question I get most from lending executives right now is about cost. Volumes have plateaued, and every budget conversation starts with “can we justify automation in a market like this?”
My response is that we’re asking the wrong question. The market you should be planning for isn’t this one. It’s the one that shows up the moment demand returns, whatever finally triggers it. Today’s market is giving lenders something they rarely have—time. The question is how we choose to use it.
I’ll acknowledge the obvious: I run operations for a company that sells automation, so you’d be right to read this with some skepticism. But the argument below isn’t about our platform. It’s about a math problem every lender in this market will face at the same time, regardless of who they buy from.
Demand in This Industry Doesn’t Rise. It Jumps.
This industry has never moved in smooth lines. Volume holds flat for long stretches, then arrives all at once, driven by whatever finally breaks the plateau: a rate move, a shift in housing supply, a change in refinance economics. The specific trigger changes each cycle. The pattern doesn’t.
Rates are the clearest recent example. We’re now four years into a rate environment above six percent on the 30-year, and anyone who has tried to predict the turn has likely been wrong more than once. That’s part of the point. If you build readiness around a specific prediction, you’re exposed every time the prediction slips. If you build it into the plateau itself, the trigger stops mattering. You’re ready whenever demand returns, not just if it returns when you guessed.
Even a gradual shift rarely produces a gradual response. Borrowers and loan officers alike watch for thresholds, not trends, and demand tends to cluster hard around round numbers and moments of visible change. A slow shift in the underlying conditions can still produce a fast spike in applications the moment one of those thresholds breaks.
The real question isn’t whether the next surge comes, or what causes it. It’s what your operation does when it arrives.
The Traditional Playbook Is Becoming Harder to Sustain
When volume spikes, the instinct is to reach for the old lever: add labor. Hire, train, absorb the surge, then quietly let the headcount shrink back down when volume normalizes again.
That playbook runs on one thing you can’t hire your way around: time. Training on a new AI or technology solution takes weeks you don’t have, and without adoption, processes begin to strain. Underwriting often feels the pressure first. It’s the function with the least room to absorb extra volume and one of the first to feel the surge. When turn times slip, a volume increase that should have been an opportunity can start costing you borrowers instead.
Automation Put in Place Now Gets Its Reps in Before You Need It
This is the part that gets missed in every “should we invest” conversation: AI-enabled automation isn’t something you “turn on” the week volume returns. It’s something you build reps into now, while volume is flat and the stakes of getting it wrong are lower. The conversation shouldn’t be, “Should we automate?” It should be, “When is the right time to automate?” My view is simple: the right time is before you need it.
This plateau is the window. I often describe it this way: automation gives operations breathing room. The work you do today is what allows your organization to absorb tomorrow’s growth without immediately adding friction. AI-enabled automation put in place now means that by the time the surge hits, your team has already worked through the exceptions, tuned the workflows, and built the muscle memory. None of that happens instantly; integration, testing, and change management take real time and effort. But that effort is far easier to absorb now, while volume is flat, than it is to attempt for the first time in the middle of a surge.
When the first 10–15% increase in volume arrives, a well-prepared operation shouldn’t feel it. That volume should flow without disrupting turn times, delaying borrowers, or forcing the team to learn a new process in the middle of a surge. That’s the buffer automation creates: meaningful capacity held in reserve without the recruiting, ramp-up, or training clock working against you.
Too often, automation ROI is measured only through labor savings. I believe that view is incomplete. The larger return is the operational capability you create—the ability to process more work, maintain quality, reduce risk, and respond faster when the market changes.
This Isn’t the Automation of a Few Years Ago
Lenders have good reason for skepticism here. Automation hasn’t historically handled the edge cases and outliers of loan origination well. That was a fair criticism of where this technology used to be. It’s a much harder criticism to make of where it is now.
The doubt that replaces that old criticism is usually about newness itself: AI is new, and new tends to mean untested. That’s true of AI in the abstract. It’s less true of AI that’s been trained and refined against years of real mortgage files, real exceptions, and real defect patterns, rather than deployed cold. The risk in “new AI” isn’t the AI. It’s the absence of that history behind it. A partner who’s spent years building and refining automation across the loan lifecycle isn’t asking you to trust an unproven model. They’re offering one that’s already absorbed the edge cases a generic model would still be learning on your file.
Many Leading Lenders Have Already Started
Knowing the technology is capable is one thing. Trusting it is another, especially in an industry that has always been slow to move, and for reasons that are hard to argue with: regulatory caution in a heavily audited business, hard-earned distrust after being burned by past vendors, the sense that the first mover always takes the biggest risk, and a reluctance to be the executive associated with job losses.
None of those concerns are unreasonable, and none of them are fully resolved just by choosing to automate. What they point to is which partner you choose, not whether to act. A partner with years of production experience in this industry, a track record instead of a pitch deck, and a genuine commitment to augmenting people rather than replacing them won’t erase every one of those concerns. But it addresses them more directly than either waiting or choosing blind.
What doesn’t go away is the cost of waiting. The lenders moving fastest, some of them as early as last year, have already put automation in place.
What they’ve bought themselves isn’t efficiency. It’s a head start on everyone still waiting to see what happens next. When demand returns all at once, a head start compounds fast, and the lenders who wait longest will be the ones scrambling hardest to catch up.
The Right Time to Build the Buffer Is Before You Need It
The best time to prepare for growth isn’t when the market turns. It’s while you still have the time to prepare thoughtfully.
When demand returns—and history suggests it will—the organizations that invested during the plateau won’t simply process more loans. They’ll do it with greater confidence, consistency, and operational resilience. That’s the real return on automation.
