Ocrolus adds document fraud detection to mortgage workflows. Ocrolus, a vertical AI workflow and analytics platform for lenders, has announced that its document fraud detection solution, Ocrolus Detect, will now be included at no additional cost for all its mortgage customers. Ocrolus Detect reviews bank statements, pay stubs, and tax forms for tampering. Underwriters see the results of the review in the Ocrolus platform or the Encompass embedded viewer where they already review the file. Ocrolus has leveraged its integration with Resistant AI to enable Detect to spot documents that are created or edited by generative AI. Unfortunately, Generative AI makes it easy and inexpensive to produce false financial documents across different types of lending. Ocrolus has determined that 12% of loan applications processed via its platform contain suspicious activity. Detect determines both what a document says and how it was created, and gives each document an authenticity score with reason codes to indicate to the underwriter where the issue was located and why it was flagged. This latter point is key. Detect identifies specific issues in flagged documents and gives lenders an explainable result. This can help lenders meet new AI governance requirements from Freddie Mac and Fannie Mae that went into effect earlier this year. To this end, Ocrolus has also integrated with Fannie Mae's Income Calculator and Freddie Mac's AIM Check, and maintains a formal AI governance program with controls mapped to both frameworks. The controls are audited every year by independent third parties and through application security testing. Lenders using Detect have already reported positive results. JMAC Lending, a wholesale lender based in Costa Mesa, California, analyzes every file in its Encompass pipeline with Detect. In doing so, JMAC prevented nearly $2 million in fraudulent loan volume in the first months of use. The lender reported that in one instance, Detect uncovered a bank statement showing a $300,000 deposit in an account that held only $1,200. "The conversation with lenders has moved from 'Can AI do this?' to 'Can I trust it enough to put it into production?'" Ocrolus General Manager of Mortgage Nadia Aziz said. "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." Founded in 2014 and headquartered in New York, Ocrolus won Best of Show in its FinovateFall appearance in 2021 and most recently demoed its technology on the Finovate stage at FinovateFall 2023. At the conference, the company showed how its AI-driven document automation platform optimizes credit decisions in small business, mortgage, and consumer lending, enabling firms to automate financial analysis and help borrowers access the financing they need faster and on the best possible terms. Ocrolus' platform analyzes 750,000 credit applications a month, and its technology powers digital lending experiences for more than 500 customers, including SoFi, PayPal, Square, and Zillow. Sam Bobley is co-founder and CEO.

Ocrolus expands document fraud detection with Resistant AI integration. 1 Oct 2026 TL;DR: Ocrolus has expanded Ocrolus Detect through an integration with Resistant AI's Documents forensic engine, combining forensic inspection of how a document was created with Ocrolus' intelligence around income documents. The expanded Detect, now live across Ocrolus' customer base, helps lenders identify tampered and AI-generated bank statements, pay stubs and W-2s with explainable, configurable risk outcomes. Ocrolus has expanded the fraud detection capabilities of Ocrolus Detect through an integration with Resistant AI's Documents forensic engine. The enhancement combines Resistant AI's best-in-class horizontal document fraud detection with Ocrolus' use-case-specific intelligence around income documents, creating the strongest document fraud detection solution on the market. The integration also directly addresses the growing threat of AI-generated documents. As AI tools make it faster and easier than ever to manufacture convincing fraudulent financial records, Ocrolus is investing more significantly in defense. The expanded Detect offering is now live across Ocrolus' customer base. Why lenders need forensic-level document fraud detection. Detect assists lenders, financial institutions and fintech and consumer platforms by identifying signs of tampering and suspicious activity in financial documents such as bank statements, pay stubs and W-2s. It is designed with scalable and flexible deployment options. The need shows up in Ocrolus' own data. Today, 12% of loan applications processed through Ocrolus contain suspicious activity. As reported by the American Bankers Association, fraud losses can account for up to 15% of annual losses for SMB lenders. Even a modest reduction in fraudulent applications funded can have a material impact on a lender's bottom line. Generative AI raises the stakes. Deepfake and AI-created documents are virtually impossible to identify through visual review or traditional checks alone. The Resistant AI integration addresses this need by combining Ocrolus' analysis with forensic inspection, including the detection of deepfake and generative AI-created documents. What the expanded Detect delivers. The expansion supports multi-document batch processing and delivers clearer, more actionable decisions that are easy to interpret. Resistant AI's forensic engine, trained on the threat patterns seen across more than 200 million global documents, detects sophisticated document manipulation techniques that complement Ocrolus' existing fraud signals. Detect now combines algorithmic signals derived from document content with deep metadata and structural analysis. That allows lenders to evaluate both what a document says and how it was created. Customers can continue using Detect's detailed Authenticity Score while also enabling simple outcome categories, such as high, medium or low risk, mapped to recommended actions. Each result is fully explainable and supported by structured flags and reason codes. Results can also be configured to match a client's specific review process and risk tolerance. "Our customers are being asked to make faster decisions while dealing with increasingly sophisticated document manipulation," said Andrew Rains, Chief Revenue Officer at Ocrolus. "By expanding Detect with Resistant AI's capabilities, we're giving customers clearer outcomes they can act on immediately to approve legitimate applications with confidence, backed by the depth, transparency and configurability they need to manage risk at scale." Built for every role in the fraud decision. The expanded Detect solution is designed to benefit multiple stakeholders, from fraud analysts seeking clear evidence and visual highlights, to underwriters who need simple, actionable verdicts, to risk and policy owners focused on precision, recall and governance. "The threat landscape has fundamentally changed. The widespread availability of high-quality forgery tools powered by AI, combined with recurring leaks of personal data, has created a new level of risk that can only be addressed by adding strong document integrity to content verification," said Martin Rehak, CEO of Resistant AI. "By combining Resistant AI's forensic engine with Ocrolus' existing strengths, we're giving lenders a defense capable of meeting the sophistication of modern fraud head-on." What Detect looks like in practice. Beacon Funding Corporation, an equipment financing company based in Northbrook, Illinois, turned to Detect as document tampering across the industry grew more sophisticated. Before this expansion, Detect was already automating document authenticity checks and flagging signs of tampering at the point of review for Beacon Funding's credit team. Beacon Funding's results include an average of $375,000 in fraudulent deals blocked per month, 30 minutes of manual investigation eliminated per flagged document and full monthly volumes maintained with a leaner credit team. Ocrolus and Resistant AI will host a live webinar, "What manual review misses: The new reality of fraud in lending," on Oct. 22, 2026, at noon ET. Speakers will cover what modern document fraud looks like, where manual review falls short and how lenders can build automated checks into the funding workflow without slowing clean deals. The session includes live audience Q&A. Register for the webinar. To learn more about Ocrolus' vertical-specific solutions that empower lenders to make faster, smarter and more profitable risk decisions, or to schedule a demo, visit ocrolus.com. Key takeaways. * Ocrolus Detect now integrates Resistant AI's Documents forensic engine, and the expanded offering is live across Ocrolus' customer base. * Today, 12% of loan applications processed through Ocrolus contain suspicious activity. * The integration adds detection of deepfake and generative AI-created documents that are virtually impossible to identify through visual review or traditional checks alone. * Detect evaluates both what a document says and how it was created by combining content-based signals with deep metadata and structural analysis. * Lenders can use the detailed Authenticity Score or simple high, medium or low risk categories, with explainable flags and reason codes configured to their risk tolerance. FAQs. What is Ocrolus Detect? Ocrolus Detect is a fraud detection solution that identifies signs of tampering and suspicious activity in financial documents such as bank statements, pay stubs and W-2s. It serves lenders, financial institutions and fintech and consumer platforms, with scalable and flexible deployment options. What does the Resistant AI integration add to Ocrolus Detect? The integration adds Resistant AI's Documents forensic engine, trained on the threat patterns seen across more than 200 million global documents. It detects sophisticated manipulation techniques, including deepfake and generative AI-created documents, and complements Ocrolus' existing fraud signals with deep metadata and structural analysis. How can lenders detect AI-generated financial documents? Deepfake and generative AI-created documents are virtually impossible to identify through visual review or traditional checks alone. According to Martin Rehak, CEO of Resistant AI, this level of risk "can only be addressed by adding strong document integrity to content verification." The expanded Ocrolus Detect combines Ocrolus' analysis with forensic inspection, allowing lenders to evaluate both what a document says and how it was created. How does Ocrolus Detect report fraud risk? Customers can use Detect's detailed Authenticity Score and can also enable simple outcome categories such as high, medium or low risk, mapped to recommended actions. Each result is explainable and supported by structured flags and reason codes. Results can be configured to a lender's review process and risk tolerance. Is the expanded Ocrolus Detect available now? Yes. The expanded Detect offering, including the Resistant AI integration, is live across Ocrolus' customer base.