Restb.ai targets UAD 3.6 compliance for lenders, appraisers

Restb.ai has released a new application programming interface called Feature UAD. The interface is aimed at helping mortgage and appraisal professionals comply with new appraisal data standards set to take effect later this year.

The tool is designed to support the Uniform Appraisal Dataset 3.6 — a revised standard that will be required for conventional loans submitted to Fannie Mae and Freddie Mac beginning Nov. 2.

Limited production under the new standard began in November 2025.

Feature UAD uses computer vision to analyze property photos and extract information such as features, condition and visible damage.

Restb.ai said the data is structured to align with the UAD 3.6 format, which significantly expands reporting requirements — including photo-level details and separate assessments of interior and exterior condition and quality.

“Our AI solution has a primary goal: to help every link in the appraisal chain speak UAD 3.6 fluently,” said Nathan Brannen, chief product officer at Restb.ai. “Whether it’s automatically identifying features, evaluating interior and exterior condition separately or detecting damages by severity and location, our outputs map to the new UAD format so clients can pre-fill appraisals, automate rules and support reviewers without adding manual work or ambiguity.”

Leaders said the structured outputs can support more than 100 automated review checks — reducing manual review and compliance risk for lenders, appraisal management companies and appraisal software providers.

The system does not generate completed appraisal reports or UAD 3.6 XML files, but instead integrates image-based analysis into existing appraisal and review platforms.

Tony Pistilli, president of valuations at Restb.ai, said the transition to UAD 3.6 requires coordination across the industry.

“Even if 90% of the industry is ready, the 10% can bring the whole system to a halt,” Pistilli said. “UAD 3.6 is like a strand of holiday lights: every stakeholder must be ready, or nothing works. Our solution helps remove the friction by equipping platforms and professionals with the data clarity they’ll need to keep the lights on.”

Restb.ai said its artificial intelligence models are trained to extract hundreds of data points from real estate and inspection photos and present them according to UAD 3.6 definitions. The new API is already being used to configure automated validation rules within appraisal platforms and review tools.

The company also cited concerns raised in its recent research about the potential for inconsistent condition and quality scoring under the new standards.

By providing photo-based assessments for interior and exterior conditions, Restb.ai said the technology can help reduce appraisal revisions and support more consistent reporting.

“This is about enabling, not replacing, appraisers,” Pistilli said. “We’re providing tools that allow appraisers to create more powerful data-driven, consistent, and reviewable reports. UAD 3.6 is a massive change, but with the right technology, it doesn’t have to be a painful one.”