Ben Anderson has a habit of saying the thing most mortgage executives will only say off the record. The loan officer who built a career on hustle, the one who answers the nine o’clock text and personally walks every file to the closing table, is on the way out. Not the person, Anderson is quick to clarify. The way that person works.
It is a provocative position, and he has the resume to make people listen. Over his career Anderson has closed more than $4 billion in mortgage volume and ranked among the top 10 loan originators in the country six separate times. He built Ben Anderson 365 into one of the better-known mortgage coaching platforms in the country and wrote a bestselling book on the path to homeownership. Today he runs Low Rate Co, which he markets without hedging as The AI Mortgage Company. The open question is whether the rest of the industry, and its regulators, are anywhere near where he is standing.
The Numbers Are Moving His Way
Adoption is climbing fast enough to make his case for him. A 2025 survey by the Stratmor Group found that 38 percent of mortgage lenders reported using artificial intelligence in 2024, more than double the 15 percent a year earlier, and Fannie Mae has projected the figure will pass half the industry. The work most exposed to automation is the work that fills an originator’s day: intake, document collection, status updates, and the endless questions about rate locks and timelines. Consulting firm EY has found that about seven in ten lenders rank cutting operating costs among their biggest challenges, and document-heavy automation is one of the clearest levers.
The largest players are already deep into it. HousingWire has reported one national lender rolling out enterprise AI across more than 600 branches and 2,000 loan officers, with executives framing it the same way Anderson does, as a way to augment people rather than cut them. At Low Rate Co, Anderson says that division of labor is already running in production.
His pitch is less about technology than arithmetic, and he tells it as a story about his own limits.
“Look, I closed loans by hand for twenty years, so I’m not knocking the hustle,” Anderson says. “I’m knocking the math. You can be the hardest-working person in your branch and still cap out around forty active files, because there are only so many hours in a day. And the forty-first client, the one who texts you on a Sunday, just quietly feels ignored. That was never a work-ethic problem. It’s an arithmetic problem, and a computer is better at arithmetic than I am.”
Where the Skeptics Dig In
Not everyone shares his timeline, and the objections are not coming from technophobes. Writing in National Mortgage Professional, industry voices have argued that lending demands consistent, explainable reasoning, the same inputs producing the same defensible decision every time, which regulators require and which probabilistic language models do not reliably guarantee. The distance between ambition and practice is also wide. Research cited by HousingWire found that while roughly 80 percent of institutions are experimenting with AI, fewer than 5 percent have moved those efforts into production.
And the hardest part of the job may be the part AI handles worst. As one analysis in Thesis Driven framed it, the messiest 20 percent of loan files tend to absorb 80 percent of the effort, and those complicated cases are exactly the ones automation cannot yet untangle. A Rocket Mortgage executive quoted in the same piece put the obstacle simply: there are too many permutations in a mortgage for any model to handle them all cleanly. The clean loans AI can breeze through were never the ones eating the labor.
Anderson’s Answer
Anderson does not wave the criticism off. He concedes more of it than you might expect.
“Do I think a machine should deny somebody a mortgage tomorrow with no human anywhere in the loop? No. The compliance people are right to be nervous, and honestly, half the stuff being sold as AI in this industry right now is a chatbot with a markup,” he says. “But that’s an argument for doing it carefully. It is not an argument for pretending it isn’t coming. Careful and inevitable are not opposites.”
“Careful and inevitable are not opposites.”
His own framing keeps the human on the decisions and hands the machine the volume. The technology, in his telling, exists to clear the repetitive load off a loan officer’s desk so the person can spend their time on judgment, strategy, and the borrower who needs talking through a hard file. That is also, not incidentally, the thing the skeptics agree AI is worst at replacing.
Who Should Actually Be Worried
Where Anderson draws a harder line is on which loan officers survive the shift. Giants like Rocket Mortgage and Better have spent years building proprietary technology and are layering AI into it. Anderson’s audience is the independent originator competing against those budgets without one, and his message to them, delivered through Ben Anderson 365, is not gentle.
“I get asked if I’m trying to put loan officers out of work, and it’s almost the opposite,” he says. “The good ones, the people who actually advise their clients, they’re about to get a superpower. The ones whose entire pitch was ‘I answer my phone fast,’ yeah, they should be nervous, because a computer answers faster and it never sleeps. That’s not me being harsh. That’s just the next few years.”
Even his critics tend to concede the direction of travel. The fight is over the timeline, and over how much of the job a machine can safely carry. On the how, Anderson is willing to be patient. On the whether, he is done waiting.
