TL;DR: Haventree Bank, a Schedule I bank specializing in alternative residential mortgage lending, cut bank statement review time by 67% after automating bank statement income analysis and fraud detection with Ocrolus. A review that once took up to three hours per file now takes 30 minutes to an hour, and the team reviews 33% more bank statements per file. The efficiency gains let Haventree Bank manage increased loan volume without adding headcount.
Haventree Bank has cut bank statement review time by 67% since adopting Ocrolus. Haventree Bank is a Schedule I bank specializing in alternative residential mortgage lending, helping non-traditional borrowers access financing through prudent underwriting and risk management. As loan volumes grew, its underwriting team needed a scalable approach to bank statement income analysis.
Haventree Bank turned to Ocrolus, a vertical AI workflow and analytics platform for lenders, to automate income analysis and scale through volume without adding headcount. Today the team reviews 33% more bank statements per file, three times faster, and completes a full 12-month review in under one hour.
Before Ocrolus, Haventree Bank’s underwriting team relied on a fully manual process for bank statement income analysis. Each file required manually transferring data from physical statements into an Excel-based income calculator to determine qualifying income.
“It was very manual before,” said Natalie Dixon, AVP of Mortgage Originations at Haventree Bank. “We were transporting all of the information from the physical bank statements into a spreadsheet and using an Excel version of a calculator… So it was a very timely and onerous process.”
Depending on file size, a single file could take up to three hours.
With Ocrolus, Haventree Bank’s underwriting team now uses automated bank statement income analysis to capture and categorize transaction data without manual transcription, giving underwriters a structured output to review rather than PDFs to manually mine.
The shift changed how the team spends its time. “Now we’re spending more time analyzing the information… So we have time to make sure that we’re capturing everything,” Dixon said.
Ocrolus Detect adds a fraud detection layer, helping the team flag potentially inauthentic bank statements before they reach the credit decision stage. “It helps us identify [inauthentic] bank statements a lot faster, so we could do some extra due diligence on them and pivot from there if need be,” Dixon said.
Since adopting Ocrolus, Haventree Bank has cut bank statement review time by 67%. Depending on file size, what once took up to three hours now takes less than one. The team now reviews a consistent 12 months of statements, where before the practice varied between nine and 12.
“Cutting bank statement review time down to between 30 minutes and an hour is a great savings for us,” Dixon said.
Accuracy has also improved. By removing manual transcription, the platform captures deposits more consistently. “The Ocrolus system captures the information accurately,” Dixon said. “I haven’t seen any use cases where it said something’s a deposit where it’s not.” Edge cases requiring manual income determination are rare, freeing underwriters to focus on higher-value analysis.
The efficiency gains have given Haventree Bank flexibility to manage volume fluctuations without growing its team. “[The automation] helps us manage the increased loan volumes without necessarily increasing headcount,” Dixon said. “We’ve been able to scale a bit, which is always a good thing.”
For Dixon, the case for Ocrolus goes beyond what it saves on each file. “Lenders should look beyond the automation itself and consider the operational impact AI has,” she said, “because it did create a lot of efficiency gains with us here at Haventree.”
As a Canadian lender working with a U.S.-based vendor, Haventree Bank has also valued Ocrolus’ willingness to adapt. Minor terminology differences between Canadian and U.S. mortgage conventions were addressed through hands-on support from the Ocrolus account team, who trained the bank’s underwriters on how to get the most from the platform within Canadian workflows.
That responsiveness has kept the relationship strong since 2019. “Our relationship has grown a lot stronger and we’re doing a lot more collaboration,” Dixon said.
“With Ocrolus, our bank statement review is consistent, accurate and three times faster,” she added. “Cutting it down to 30 minutes to an hour is a great savings for us.”
Read the full Haventree Bank customer story, or see how Ocrolus supports mortgage lenders.
Haventree Bank uses Ocrolus to automate bank statement income analysis and fraud detection. Ocrolus captures and categorizes transaction data without manual transcription, and Ocrolus Detect flags potentially inauthentic bank statements before they reach the credit decision stage.
Haventree Bank cut bank statement review time by 67%. Depending on file size, a review that once took up to three hours now takes 30 minutes to an hour.
Haventree Bank now reviews a consistent 12 months of bank statements per file. Before Ocrolus, the practice varied between nine and 12 months.
Detect adds a fraud detection layer that helps Haventree Bank identify potentially inauthentic bank statements faster. The team can then do extra due diligence on those statements before the credit decision.
No. The efficiency gains from automation have helped Haventree Bank manage increased loan volumes without increasing headcount.