Stephen Butler on AI’s next chapter: Clean data, background workers, and manufacturing-level mortgage automation

True’s new CEO shares why eliminating stare-and-compare is only the beginning

In this conversation with HousingWire’s Allison LaForgia, TRUE CEO Stephen Butler discussed how decades of experience in process optimization across a wide range of industries refined his vision for the mortgage space and TRUE’s evolving role in it.

“The common theme in my career is business process optimization,” Butler said. “And while I’ve only been in mortgage for about 10 years, I’ve seen the same issue here as in other complex industries — very few people understand the entire process from end to end. That’s a huge opportunity for change.”

TRUE, which has been at the forefront of mortgage AI for a decade, isn’t interested in following the latest AI trends. It’s defining them. At the center of that strategy is what Butler calls “background AI workers,” which go beyond assistant tools to fully handle specific tasks.

“If you have to deal with the output of a tool manually, are you really getting a lift?” he asked. “That’s the problem with assistant-based AI. It still requires human cleanup. But when AI can fully manage a manufacturing step in the mortgage process, the ROI is real.”

Start with clean data

For those wondering where to begin, Butler’s answer is simple: focus on data. “We’re strong advocates of getting your data right first. That means data that’s full, complete, and consistent — across documents and systems,” he said. “If you can do that early in the loan process, your decisioning—pre-approvals, income and asset calculations — can start almost immediately.”

The industry’s overreliance on manual verification, what Butler called “stare and compare,” is a major drag on productivity. “The poor underwriter has to close the file and kick it back to the LO or processor to fix mismatches. It’s slow and inefficient,” Butler said. “AI can do it. The right AI can eliminate that step entirely.”

A practical roadmap, not a moonshot

TRUE has found the highest rate of success comes when you help clients tackle AI adoption one step at a time. “No one’s trying to boil the ocean,” Butler said. “The lenders we’re working with — big national lenders — are rolling out background AI workers thoughtfully. They’re seeing lift in LO productivity, faster pre-approvals, and better customer experiences.”

“The LOs now have more time for white-glove service, and the processors can double or triple their productivity.”

What’s next for True

Butler said their AI’s accuracy has reached new heights. “We’re seeing results better than human accuracy using machines only. Our solution doesn’t have to be offshored for review — we handle it all in the background.”

True is also launching integrations with underwriting platforms and expanding its Mortgage Operation Service (MOS). “We think of this as manufacturing, like physical products — downhill instead of back and forth. Once we nail that, we’re going to see much faster loan production.”

As Butler put it, “We’ve probably run tens of millions of loans through the system. Now it’s about bringing that horsepower to every step of the mortgage manufacturing process.”