30-Yr. Fixed Conforming. Updated hourly during market hours.
Single family homes on the market. Updated weekly.Powered by Altos Research
Market yield on U.S. treasury securities at 10-year constant maturity. Updated daily.Data sourced from FRED

The future of mortgage lending: Rafael Goldberg on AI decisioning, compliance and competitive advantage 

Sapiens AI isn’t just hype – it’s a game-changer for mortgage lenders. Diego Sanchez has Rafael  “Rafi” Goldberg of Sapiens to discuss the evolution of AI decisioning and its role in modernizing mortgage operations. He details why AI is no longer optional and how lenders can harness its power without major tech disruptions. 

This conversation delves into what AI decisioning means in practice, from reducing cycle times and costs per loan to enhancing compliance and building stronger relationships with borrowers. Rafael also shares a forward-looking perspective on how AI will continue to reshape mortgage lending, giving early adopters a significant competitive edge. 

“We have an AI decisioning platform,” Goldberg explained. “It’s really a new category bringing together three evolving disciplines — machine learning, business rules and decision management, and data analytics — into one cohesive experience. Each has operated in silos for years. We’re unifying them so businesses can engage with the entire ecosystem seamlessly.”

The company didn’t stumble into mortgage. It was invited. “We got our start in business rules and rapid application development,” Goldberg said. “One of the GSEs loved what our model did for their business and asked us to build an enterprise platform around it. That’s how Sapiens Decision was born. We’ve since worked with major lenders, banks, and insurers, and now we’re reinvesting in our mortgage roots.”

Asked what AI decisioning actually looks like in practice, Goldberg described a feedback loop between people, policy, and data. “At the core are decision models — representations of business policy. Generative AI helps you build those structures, but you don’t go straight to production. You keep humans in the loop with testing and validation,” he said. “Then you execute those models, gather analytics, and apply AI again to introspect that data — seeing how to improve outcomes. It becomes a virtuous cycle.”

Sanchez asked whether lenders need to overhaul their existing tech stacks to integrate. “No ripping and replacing,” Goldberg said. “We’re API-first. Our platform sits above your systems, elevating decisions as a first-class enterprise asset. Everyone can point to where data lives or where customer experience lives—but where do decisions live? We make that visible and governable.”

When it comes to users, Goldberg said the company bridges the gap between business and IT. “It’s a marriage between policy and execution,” he said. “We empower functional leaders to define business logic clearly, and we empower developers to build without interpreting ambiguous requirements.”

For most lenders, the draw is both speed to market and cost reduction. “Clients want to roll out new, complex products — like non-QM — faster,” Goldberg said. “This approach gets them to market quickly and cuts costs by automating what would otherwise require manual exception handling and check-the-checker processes.”

“We use AI to speed up the process, not to be the process,” he emphasized. “Our decision models are a clear box, not a black box. When an auditor comes in, lenders can push a button and show exactly why each decision was made. That consistency builds borrower trust.”

Compliance, he added, is baked in. “Every asset has an approval lifecycle—customizable, reviewable, and fully transparent. Any AI that informs a decision model is validated, regression-tested, and approved before deployment,” he said. “We’re pulling back the curtain on what most call black-box AI.”

Looking ahead, Goldberg sees the next wave coming fast. “We’re entering the agentic era,” he said. “The idea of a semi-autonomous enterprise is here. We’re already running POCs that use agentic AI to evaluate decision models against KPIs. If there’s drift, the system suggests improvements automatically. It’s game-changing.”

He paused, then smiled. “Every lender wants to hit their KPIs. Now AI can help them see exactly how.” 

To learn more about Sapiens….