customer stories
JMAC Lending stops nearly $2 million in fraudulent loans with Ocrolus Detect
Customer Stories
JMAC Lending stops nearly $2 million in fraudulent loans with Ocrolus Detect
Mortgage | Detect, Analyze and Classify
JMAC Lending is a wholesale lender based in Costa Mesa, California that works with mortgage brokers and correspondents in more than 39 states. Now in its 30th year, JMAC originates the full range of agency, jumbo, FHA, VA and USDA products alongside non-QM programs including bank statement and DSCR loans.
JMAC turned to Ocrolus, a vertical AI workflow and analytics platform for lenders, to bring consistency and speed to document fraud detection across its agency and non-agency underwriting teams. Detect now runs on every file that enters the pipeline, and JMAC has already stopped nearly $2 million in fraudulent loan volume.
Challenge
JMAC Lending's underwriters were authenticating income and bank documents by eye, a manual review that could run from 15 minutes to well over an hour per file and varied from reviewer to reviewer.
Solution
Ocrolus Detect runs automatically on every submission and every resubmission inside Encompass, giving underwriters document flags and authenticity scores before they ever open the file.
Result
• Nearly $2 million in fraud stopped across four flagged cases to date
• Manual document authentication eliminated, saving 15 to 60 minutes per loan file
• Document integrity verification running on 100% of files
• Consistent fraud review across the team, with a defined escalation and QC path
Challenge: Fraud detection rested on the eyes of the reviewer
Before Ocrolus, document authentication at JMAC came down to what a seasoned underwriter could catch on screen, which meant results varied from reviewer to reviewer. There was no consistent threshold for what counted as a suspicious document, and no reliable way to tell whether a file had been altered with software.
"What it honestly came down to was the eyes of the person reviewing the file," said Dana Reminder, Director of Underwriting at JMAC Lending. "You were really hinging it on a seasoned underwriter opening up a file, looking at a bank statement and seeing something that looked off."
The manual approach carried a real time cost as well. A standard file took 15 to 20 minutes to work through a checklist, and bank statement programs took far longer. A 24-month file could run past a thousand pages, all of it reconciled by hand.
"In this day and age, you can digitally edit anything," Reminder said. "Reconciling by hand is painful and it can take a lot of time when you are looking to streamline your process."
Solution: Detect runs on every file, inside Encompass
JMAC embedded Ocrolus Detect directly into its Encompass workflow rather than running it as a separate check. Every submission and resubmission is screened automatically, with the system watching for bank statements, paystubs and W-2s as they come in. Coverage became universal instead of selective, and no file moves forward without being checked.
"All submissions will go through Ocrolus Detect," Reminder said. "It will catch fraud at all stages."
The review itself is fast. Underwriters open the file, start with the flags, then work into the documents and authenticity scores behind them. Detect analyzes both the image and the metadata when it is available, surfacing edits that would never be visible on screen.
"The underwriter is that human in the loop that starts looking at the documentation and taking in the flags and Detect signals to truly see if something looks off," Reminder said.
From there, JMAC follows a defined path. Suspicious files escalate to a supervisor or manager, then to QC for reverification directly with the employer or the bank. Running Detect inside Encompass means every reviewer works to the same standard rather than deciding on their own when a file is worth a closer look.
"It's not enough to have a suspicion of something nefarious going on. You want to prove it, because then you act on what you can prove," Reminder said. "We have zero tolerance for fraud."
Result: Nearly $2 million in fraudulent loans stopped
JMAC is early in its use of Detect and has already identified four cases of suspicious activity. In one, flags for data edits and unreconciled balances led underwriters to a bank statement that had been altered before submission.
"We had a borrower give themselves a deposit of 300,000, and it wasn't there," Reminder said. "We validated the balance, and it was a $1,200 balance in the account."
Across the cases caught so far, the loan volume totals nearly $2 million. The financial exposure is only part of what JMAC protected. Stopping fraudulent files before they are originated keeps the company in good standing with its investors and protects the designations it depends on.
"It's your reputation, it's your designations; if it's your FHA loans, it keeps you off of being on the neighborhood watch," Reminder said. "We're not putting ourselves at risk of being disqualified from any of our investors."
On the efficiency side, Detect removes the manual authentication pass entirely, whether that step took 15 minutes on a standard file or more than an hour on a complex one. Underwriters get time back to make decisions, and the evidence to back up what they suspected.
"It will save them at least that 15 minutes that they were originally spending, up to the hours that they could be in," Reminder said. "It's saving us from originating loans that we would not originate."
Key outcomes:
- Nearly $2 million in fraudulent loan volume caught in the first months of use
- Four cases of suspicious activity identified and stopped to date
- Manual document authentication eliminated, saving 15 to 60 minutes per loan file
- Document integrity verification running on 100% of files
- One consistent fraud threshold and escalation path across the underwriting team
- Investor eligibility and FHA designations protected by stopping files before origination
Why JMAC Lending recommends Ocrolus
Detect is one part of how JMAC uses Ocrolus. The team also runs income and bank statement calculations through the platform and uses Ocrolus to index documents inside the Encompass eFolder. Reminder sees the value compounding as those layers connect, particularly as more borrower data becomes available across documents.
"As you start to wrap Detect around everything, we are going to see some exceptional change that will really give a lift to the user," she said.
The partnership behind the product has reinforced that confidence. Questions sent to Ocrolus support come back quickly, and the team keeps JMAC informed even when a file needs manual attention.
"For me it is the relationship with our account manager, and the response time when we send things into the support desk is phenomenal," Reminder said.
"We have zero tolerance for fraud. Ocrolus Detect is a partner to our underwriters to be their second set of eyes. You're not replacing them. You're enhancing them."
Dana Reminder, Director of Underwriting at JMAC Lending
