American Federal cuts complex income reviews 90% with Ocrolus
Founded in 1987 in Clifton, New Jersey, American Federal Mortgage Corporation (American Federal) has grown into one of the nation's largest privately held mortgage bankers.
American Federal partnered with Ocrolus, an AI-powered workflow and analytics platform, where a range of staff rely on Ocrolus for everything from document classification and income analysis to Inspect for discrepancy detection and condition management.
American Federal needed a faster, more consistent way to prep, review and condition loan files across its processing and underwriting teams without adding headcount to absorb the mortgage market's volume swings.
American Federal deployed the full Ocrolus suite across the file lifecycle, automating document classification and income analysis while using Inspect and conditions to surface defects early, before a file ever reaches the underwriter.
β’ Complex self-employed income reviews cut 90% β from ~1 hour to 5β10 minutes
β’ Document prep cut 67% β from 15 minutes to about 5 β through automated classification at intake
β’ Initial processor review on a typical file reduced by up to 50% with Inspect
β’ Underwriters clear higher file volumes in about nine hours instead of 12
β’ Borrower requests consolidated into a single upfront ask
Before Ocrolus, every file began with manual work. Processors and underwriters opened each loan and sorted, renamed and organized documents by hand before any analysis could start.
"Before, we had to manually rename and sort all documents," said Maya Matejko, Assistant Vice President of Operations at American Federal, work that added at least 10 to 15 minutes to every file.
Income was the heaviest lift. For self-employed borrowers, underwriters reviewed a year or two of statements line by line.
"I would have to look at 12 or 24 months of bank statements. Every transaction. It could take forever," said Gail Stibitz, Senior Vice President of Underwriting at American Federal, often close to an hour of work before the math even began.
And because issues could sit deep in a file, discrepancies often surfaced late, forcing the team back to borrowers for one more document right before closing.
"Before Ocrolus it was definitely more of a manual inspection process," said Caroline Cherry, a processor at American Federal, combing through asset statements, bank statements and pay stubs by hand.
American Federal adopted Ocrolus as an integrated workflow, introducing each capability in sequence so staff could master it before moving to the next.
Classification removed the sorting bottleneck first. "The documents are already sorted once they get in there," Matejko said. With files arriving organized and consistently named, the team could open a loan and start analyzing instead of administrating.
Income analysis was the breakthrough for underwriting. Rather than reading a year or two of statements line by line, underwriters now get a calculated income figure from Ocrolus in minutes. "Ocrolus just does that for me β it scans it all and gives me an income figure," Stibitz said
For processors, the Inspect insights changed where the work starts. "I check right off the bat what items it's identifying that we don't have yet, which saves me time as opposed to combing through the file myself," Cherry said.
Conditions now push that work even earlier. "Before the file even hits underwriting, the processors are able to add conditions and obtain documentation," Stibitz said. "By the time it hits underwriting, we're not having to look and ask for it."
The compounding effect across roles has been significant and measurable.
Classification saves time before review even begins. "The indexing alone saves a ton of time," Stibitz said. On income, the gain is dramatic: a complex self-employed file that once took about an hour now takes 5 to 10 minutes.
Most strategically, the efficiency lets American Federal hold its team steady through the market's surges. "When it gets busy, an underwriter can reach 70, 80, 90 files. We get that work done; instead of it being 12 hours, it's 9," Stibitz said, capacity that protects the company's people rather than its margins.
Inspect and conditions have also reduced borrower friction. By surfacing requests early, the team consolidates what used to be multiple follow-up calls into one upfront ask. "It's a much better experience when we have those conditions and requests all in one spot at the very beginning," Cherry said.
Across processing, underwriting and leadership, the team's advice converges on one idea: trust the tool, and keep the human in the loop.
"Lean into it and trust the system," Cherry said. "A lot of the time it's catching things we don't see." Stibitz frames it the same way for underwriters: "Don't be afraid to use it as a double tap. You're not taking away the human experience, you're making sure it didn't miss anything."
The partnership itself has reinforced that confidence. "Ocrolus has always been not only open, but encouraging us to provide feedback, and they've implemented a lot of the things we've discussed," Stibitz said.
"Ocrolus scans documents from complex self-employed income reviews and gives me an income figure. Itβs wonderful. It takes me from 1 hour down to maybe 5 or 10 minutes."
Gail Stibitz, Senior Vice President of Underwriting at American Federal Mortgage Corporation