The real estate valuation industry is on the horizon of a major transformation in decades with the rollout of UAD 3.6 standards and the continuously evolving AI landscape. It is an opportunity to modernize how appraisals are performed, captured, exchanged, and analyzed. AI is becoming an assistive technology for completing appraisals, reviewing the reports, extracting insights from property images, automating routine workflows and helping appraisers focus on higher-value decisions.

While that sounds good, it also raises many questions. How will appraisers adapt to the new standards and technologies? Will these capabilities improve the quality of the appraisals and reduce the risk in the valuations? How will this change the turn around times for appraisals and lenders’ processing times? How will all this benefit the borrower or the homeowner – the ultimate customer?

Here are a few realities

• Appraisers feel that UAD 3.6 is converting them into data gatherers, forcing them to catalog granular property features that ultimately feed AI valuation models.
• Turn around times will most certainly increase in the first few months as the change runs its course across the industry. Automation of data capture and workflows will eventually help as it gets adapted.
• Regardless of the automated mechanisms used to capture the property data, the appraiser is strictly liable for the entire report.
• When automated quality control works through the rigid machine readable data captured in the appraisal reports, it can create a wave of hardstops requiring hours of writing from the appraiser to provide narrative commentaries to override them.

Our Strategic Approach

Multimodal Data Collection: Relying on any one source of data is not sufficient to ensure capturing reliable property data. Advanced valuation systems combine “reliable” computer vision models, MLS and assessor data, and geospatial analytics to arrive at a high-confidence property data.

Model-Driven Valuation: While UAD 3.6 increases the data to be gathered and modifies its representation, it is important to note that the fundamentals of valuation standards – USPAP and Appraisal Methodologies are not changing. So, it is important for valuation systems to stay grounded in those methods to ensure we do not increase the core risk in property valuations due to a new data standard.

Pre-Submission AI Quality Control: Automated QC checks embedded in valuation systems perform a comprehensive QC of the appraisal specific to a lender’s LOE or the GSE’s rules. This will drastically reduce the revision rates and improve the overall quality and turn times.

Understanding Property Complexity: The valuation system should understand the complexity of a valuation to provide data reliability guidance to the appraiser. Computer vision outputs could be questionable for luxury or custom homes or a rural parcel with unique layouts. A computer vision data reliability score (or a confidence score) would alert the appraiser and reduce rework.

And here’s our UAD 3.6 Update

We are fortunate to have been serving this industry for 10+ years and be part of this transformation.

We started our UAD 3.6 journey with quality control. Profet Review is ready to perform a fully automated QC check for UAD 3.6 zip files. This includes our proven models: computer vision to validate condition, quality, and property features, language processing for bias, and comparable analysis.

Profet Valuation is being upgraded for UAD 3.6 data capture with conditionality and validation capabilities. The system that currently serves various alternate valuations and current UAD forms is built on the concepts of model-driven valuation. So, UAD 3.6 is another data gathering and output document within the system. While we may be a little late to the UAD 3.6 party with our valuation product, we look forward to serving the industry with all the strategic solutions presented above.