Ocrolus, a vertical AI workflow and analytics platform for lenders, announced that every Ocrolus mortgage customer now receives Ocrolus Detect, the company's document fraud detection product, at no added cost. The move makes fraud screening a standard part of the mortgage workflow rather than a paid add-on. Detect checks bank statements, pay stubs, and tax forms for tampering, and underwriters can view the results directly in the Ocrolus platform or in the embedded viewer inside Encompass, the loan origination system where they already review loan files.

Lenders are already running Detect in production. JMAC Lending, a wholesale lender based in Costa Mesa, California, screens every file in its Encompass pipeline with the tool. In its first months of use, JMAC stopped nearly $2 million in fraudulent loan volume. In one case, Detect led underwriters to a bank statement showing a $300,000 deposit in an account that held just $1,200.

The rollout comes as generative AI has made fake financial documents cheap and fast to produce. Across all lending types, 12% of loan applications processed through Ocrolus contain suspicious activity, and a single altered document can trigger a repurchase demand, under which the lender must buy the loan back from the investor, or the loss of a designation.

“The conversation with has moved from 'Can AI do this?' to 'Can I trust it enough to put it into production?'” said Nadia Aziz, General Manager of Mortgage at Ocrolus. “Fraudulent bank statements and income documents can now be created in minutes and are becoming harder to distinguish from the real thing. Fraud screening should be built in, not an added expense for lenders. Detect shows underwriters exactly why a document was flagged, and the underwriter makes the call.”

That trust starts with evidence. Through Ocrolus' integration with Resistant AI, Detect now catches documents that generative AI created or edited, examining both what a document says and how it was made. Each document receives an authenticity score with reason codes that show the underwriter what looks wrong, such as edited text, balances that do not reconcile, or changes made to the file after it was created.

That transparency now carries regulatory weight. Freddie Mac's and Fannie Mae's AI governance requirements — rules the two government-sponsored enterprises apply to the lenders that sell them mortgages — took effect earlier this year, and every Detect flag provides its reasoning, giving lenders an explainable result. The same approach underpins Ocrolus' integrations with Fannie Mae's Income Calculator and Freddie Mac's AIM Check. Behind the product, Ocrolus maintains a formal AI governance program, with controls mapped to both frameworks to help meet the obligations they set. Those defined controls are audited annually by independent third parties under ISO 27001, SOC 2, and PCI DSS, along with application security testing.

Putting AI into production also takes people who know the work. Ocrolus' in-house mortgage underwriting experts bring firsthand lending experience to product design and accuracy, and partner with account managers to support each rollout through onboarding, onsite training, and ongoing adoption.

That combination is driving lenders to entrust Ocrolus with more of the workflow. In the past 90 days, 40 lenders signed up for Ocrolus Inspect, a product that identifies data discrepancies and automates underwriting conditions.

Rebecca Seward, Vice President of Product at Ocrolus, is scheduled to demo the full Ocrolus mortgage suite on the Hub Stage on Monday, Oct. 12 at the MBA Annual Convention and Expo in Chicago.