Two homes in the same neighborhood can have the same square footage, a similar floor plan, and very different sale prices. One has an updated kitchen and well-maintained bathrooms. The other shows visible wear and deferred maintenance. If the data available to an automated valuation model (AVM) describes those homes in much the same way, the model has less information to explain why buyers might pay more for one.
That is the problem Veros Real Estate Solutions (Veros®) set out to address when it launched VeroVISION on June 3, 2026. VeroVISION uses AI-powered computer vision to turn property photos into condition data. In VeroVALUE Elite, that data helps the AVM account for differences between homes that can look similar in conventional property records.
In Veros launch testing against VeroVALUE, VeroVALUE Elite’s P10 accuracy improved by more than 5% for homes in poor condition. P10 measures the share of AVM estimates that fall within 10% above or below a benchmark value, such as the sale price.
How VeroVISION turns photos into condition data
VeroVISION identifies what each photo shows, such as a kitchen, bathroom, or exterior, and evaluates observable signs of condition, including visible wear and damage. It combines individual image scores into room-level scores and uses those results to produce an overall Veros Home Score (VHS) on a condition scale from 1.0 to 5.0. Higher scores indicate better property condition.
That level of detail can be useful when two properties share the same standardized appraisal condition rating. For example, two homes rated C4 may both show moderate wear, yet the photos may show that one has better-maintained bathrooms and less wear on its flooring. Room-level scores give reviewers another way to compare those homes and identify differences worth a closer look.
In Veros testing, the VHS showed a 93% correlation with condition assessments from real estate professionals. Veros reports score coverage for approximately 97% of U.S. residential properties, including scores estimated from other data when photos are unavailable.
What changes when an AVM accounts for condition?
VeroVALUE Elite combines VeroVISION insights from available property imagery with sales and local market evidence. When available, a condition adjustment estimates the premium or discount associated with the home’s condition relative to its local market. That local context matters because buyers in different markets can respond differently to the same level of wear or upkeep.
In launch testing against the VeroVALUE AVM, Veros reported more than a 3% improvement in nationwide P10 accuracy for VeroVALUE Elite.
Where condition data can help beyond an AVM
VeroVISION is also available as a standalone service. Users can request a VHS by property address, giving lenders, servicers, and investors another input when deciding where to focus their review.
For a lending portfolio team, that could mean flagging properties with condition scores that warrant a closer look. For an appraisal reviewer, it could mean examining a comparable sale more closely when the available photos suggest a meaningful condition difference. In both cases, the score helps direct attention, and the reviewer can weigh it alongside the other evidence in the file.
What is next for property condition data?
One opportunity is to bring structured condition data into workflows that already use property photos or virtual inspections. A reviewer could use room-level observations to see where wear is concentrated and decide which areas need more documentation. In a virtual inspection workflow, an appraiser could review the system’s room-level observations, correct them where needed, and weigh them alongside other evidence when developing an opinion of value.
Explore VeroVISION or see how condition data informs the VeroVALUE Elite AVM.
About Michael Tokar, Marketing Specialist
Michael Tokar currently attends Illinois State University, and studies Business Marketing.








