Mortgage lender Newrez is set to deploy fintech Valon’s AI-driven mortgage servicing tool to more than 4 million homeowners.  

Valon’s AI-native platform, which automates and centralizes mortgage servicing processes as a single “source of truth,” was designed to address fragmented data across systems, Chief Executive Andrew Wang told FinAi News. Some mortgage servicers operate on dozens or hundreds of systems, he added. 

“When systems are that complex with so many different pieces, they inevitably don’t really talk to each other that well,” he said. “And that drives inconsistency, data problems, homeowner dissatisfaction.” 

Homeowner POV of ValonOS 

Newrez is especially optimistic about the Valon platform’s ability to observe user operating behaviors and generate agentic AI–driven automations, Chief Information Officer Brian Woodring told FinAi News.  

“Because this platform is AI native and incorporates built-in observability across each step of the process, it significantly streamlines the creation of accurate, effective agentic AI automations,” he said. “This will lead to a faster, more intuitive customer experience while reducing our cost-to-service.”  

Newrez is the third-largest mortgage servicer and the fifth-largest mortgage lender in the U.S., according to Newrez parent company Rithm Capital. 

Rithm this month announced it would expand its equity stake in Valon.  

Newrez’s servicing portfolio totaled $852 billion in unpaid principal balances at the end of 2025, up 1% year over year, according to Rithm’s Feb. 6 earnings release. Its originations increased 7% YoY in 2025 to $63.3 billion.  

4x productivity gains 

Mortgage servicing employees using Valon’s operating system, ValonOS, can manage approximately 3,000 loans at a time, compared with an industry average of about 750, Wang said.  

To this end, Newrez expects a substantial increase in accounts managed per team member as the system automates a significant portion of “behind-the-scenes work, particularly back-office tasks that are often manual and time intensive,” Woodring said. “This will enable our operators and customer-facing agents to devote more time” to customer support, he added. 

Homeowner POV of ValonOS

Newrez also expects meaningful gains from expanded digital self-service capabilities through conversational AI, “allowing customers to resolve many of their needs independently without requiring direct human assistance,” he said.  

The company will transition to ValonOS in 2027.  

Built for compliance  

Mortgage servicing is the administration of a home loan after closing, including collection of monthly payments, managing escrow accounts, customer communication and loan modifications. 

Building an AI-native system that can handle the highly regulated nature of mortgage servicing was imperative for Valon, Wang said.  

“The regulation, to me, is both the reason the problem exists and the reason why there is opportunity.”

— Andrew Wang, CEO, Valon

Many servicers have layered systems on top of each other to address increasing regulations, particularly after the housing crisis of 2008, he added.

A strong data infrastructure, explainability and reinforcement learning are key to building AI agents that can meet compliance requirements, Wang said. 

Further, the fintech is building AI agents to perform data migrations from old systems to the ValonOS, Wang said. Valon is also working to capitalize on the growing capabilities of voice agents, he said. 

Valon, founded in 2019, had raised approximately $230 million after its latest funding round in 2024, with investors including Westcap, New Residential Investment, Kairos and Andreesen Horowitz, according to a company release. 

Register here for the inaugural FinAi Banking Summit, taking place March 2-3 in Denver. View the full event agenda here.