There’s a significant blind spot in the real estate appraisal process — one that could be costing lenders over $27 billion annually in repurchase risk. That’s according to a new white paper from AI-driven computer vision firm Restb.ai, “The Impact of Condition and Quality on Appraisal Accuracy.”
Appraisers traditionally assess a property's condition and quality using the Uniform Appraisal Dataset (UAD) 6-point scale. But that method often leads to clustering of homes in the middle categories — 86.1% of properties were rated as C3 or C4 for condition, and 97.0% as Q3 or Q4 for quality, Restb.ai found.
Such “clustering” can obscure meaningful differences between properties. Further, inconsistencies arise when adjustments are made even when both the subject and comparable properties share the same condition or quality ratings.
The study found that appraisers still applied condition adjustments in 11.8% of such cases and quality adjustments in 5.3%, raising questions about the consistency and transparency of these evaluations.
Financial Implications: A $27 Billion Risk
By analyzing 1,271 appraisals and 6,495 comparable properties, Restb.ai identified that 33.6% of appraisals had a “high risk” of inadequate or missing adjustments, leading to potential repurchase risks exceeding $27 billion annually. What’s more, 73.9% of appraisals exhibited a medium risk, suggesting that the issue is actually more widespread and systemic.
These inaccuracies can result in overvaluation or undervaluation of properties, affecting borrower equity and increasing the likelihood of loan repurchases, which are costly for lenders. The integration of AI into the appraisal process can serve as a quality control mechanism, flagging potential issues early and reducing financial exposure.
AI’s Objective, Granular Analysis
Unlike human appraisers, AI evaluates each property independently, free from regional biases or subconscious influences. It provides decimal-level scoring, allowing for more nuanced distinctions between properties.
Additionally, Restb.ai's computer vision tech evaluates individual components of a home — such as the kitchen, bathrooms, interior, and exterior — offering a detailed understanding of renovations and updates that might not be captured in traditional assessments.
The white paper highlights a need for modernizing the home appraisal process. By incorporating AI-driven evaluations, lenders and appraisers can achieve more accurate, consistent, and objective assessments of property condition and quality, the company argues — and doing so would not only mitigate financial risk but promote fairness and transparency in real estate transactions.
