Verus Mortgage Capital has replaced its legacy loan origination system with Vesta’s platform, deploying new automation and artificial intelligence tools across its Non-QM operations.
Verus purchases non-agency loans from correspondent lenders in all 50 states and the District of Columbia. The company said it selected Vesta partly because the system can accommodate proprietary lending guidelines and the exception-driven workflows common in Non-QM lending.
“Our business is built on flexibility and expertise across a wide range of borrower scenarios. We needed an LOS that could match that,” said Dane Smith, president of Verus Mortgage Capital. “Vesta gives us a platform that scales with us across channels and creates opportunities to apply automation and AI in ways that improve efficiency and the experience we deliver to our lending partners.”
Vesta structures the origination process as a series of individual tasks that can be assigned to employees or AI agents. According to the companies, three AI-based features are now operating within Verus’ production environment.
Document Intelligence identifies and processes loan documents, while an AI agent can complete certain operational tasks. A separate Loan Assistant allows users to ask questions about a file or make changes using plain-language commands.
“Non-QM is one of the hardest origination problems in mortgage: the rules are proprietary, exceptions are the norm, and the workflow has to flex constantly. That’s exactly why Verus is on Vesta,” said Mike Yu, co-founder and CEO of Vesta. “Our platform was built to handle that complexity natively, and our AI agent can do the operational work behind it.”
Vesta says its agents can perform tasks ranging from application reviews and employment verification to credit analysis, initial underwriting decisions, condition clearing, and closing-package reviews. Lenders determine which actions an agent can take independently and when a file must be escalated to an employee, according to Vesta. The platform also logs the agent’s actions for review and auditing. The companies did not disclose performance results from Verus’ implementation.
Jeffrey Pisano, head of information technology at Invictus Capital Partners, which backs Verus, contrasted the implementation with AI tools added to older origination platforms.
“We evaluated what AI looks like inside legacy LOS platforms — bolted-on assistants that read documents and make suggestions — and what it looks like inside Vesta. There’s no comparison,” Pisano said. “Vesta’s agents can read our guidelines, complete real underwriting tasks, and operate inside the same system our team uses every day: the only architecture that works for Non-QM, where the rules are ours and the work is judgment-heavy.”
Vesta says its platform has helped some lenders reduce operational costs by as much as 25%. Neither company said Verus has achieved that level of savings.
The system change is part of a broader investment by Non-QM companies in technology designed for proprietary loan programs rather than agency production. As the sector grows, lenders and investors are looking for ways to process more complicated borrower files without allowing manual work and exception reviews to become operational bottlenecks.
*This article was primarily written by a human author. AI tools were used in a limited capacity for research assistance or light editing.
