For years, the mortgage industry has worked toward a simple objective: make valuation faster. And for good reason. Technology has transformed how lenders, servicers, investors and other housing finance organizations obtain property and market information. Automated valuation models (AVMs), digital property data, evaluations, desktop valuations, inspections and other technology-enabled solutions can deliver insight far faster than traditional processes alone. That speed has real value, particularly in an industry where efficiency, cost and the ability to scale matter every day.
But as valuation becomes faster and more automated, another question deserves more attention: How do you know when a valuation is strong enough to trust, and when it deserves a closer look?
That is the Valuation Trust Gap. It isn’t an argument against AVMs or automation. Quite the opposite. The more valuation decisions an organization can make efficiently through technology, the more important it becomes to understand the information supporting those decisions and recognize situations where additional context may matter. The goal isn’t to slow valuation down. It is to become smarter about where additional diligence is warranted.
A Valuation Is More Than a Number
A valuation is an estimate and/or opinion, and the quality of that estimate depends on the data, methodology, market conditions and property characteristics that inform it. Modern AVMs can process enormous amounts of information and apply sophisticated modeling techniques to estimate a property’s value. They can provide value estimates, and other measures that help users understand the result and its expected reliability. VeroVALUE, for example, combines predictive modeling, quality data, property information and local market data and provides confidence scores alongside its valuation results.
But even sophisticated valuation models operate within the information available to them. That distinction becomes particularly important in real estate because markets are local and properties are not interchangeable. Two homes that appear similar based on square footage, bedroom and bathroom counts, lot size and location may have very different market characteristics. One may have been substantially renovated while another has significant deferred maintenance. Those differences can affect marketability and value, even when the properties look similar in the data.
That doesn’t mean an automated valuation has failed. It means property context matters, and valuation technology continues to evolve to capture more of that context.
The Next Question for AVMs
Property condition is a good example. It is a property-specific characteristic that can be difficult to capture consistently and incorporate into an automated valuation. Newer approaches are changing that equation.
VeroVALUE Elite, for example, incorporates AI-driven property condition scoring and computer-vision intelligence into the valuation process. By analyzing property photos and incorporating condition intelligence into the model, VeroVALUE Elite adds another dimension of property-specific information to the valuation.
That evolution is important because it changes the conversation. The question is no longer simply whether an AVM can produce a value quickly. Increasingly, the question is how much relevant property context can be incorporated into that valuation and how that information can help users make better decisions.
That does not mean condition information eliminates uncertainty. No valuation methodology can eliminate all uncertainty from a dynamic housing market. It means the valuation can be informed by additional information that may be relevant to the property being valued.
Confidence Is Not Certainty
There is another distinction worth making as valuation technology becomes more sophisticated: confidence is not the same thing as certainty. Confidence and other model performance measures can provide valuable information about a valuation, but they need to be understood within the context of the model, the property and the data available. A confidence measure can be useful, but it shouldn’t be viewed in isolation. The broader question is whether the available data and property characteristics provide enough support for the decision being made.
That might mean considering whether the property has characteristics that make it less comparable to nearby sales, whether the market is changing faster than historical patterns suggest, whether the available property information is sufficiently current, or whether condition appears consistent with the information supporting the valuation. It could also mean looking at whether multiple valuation methods are producing results that tell a consistent story.
The point is not to second-guess every automated valuation. That would defeat much of the value technology brings to the process. The point is to recognize that not every property presents the same level of uncertainty, and therefore not every property needs the same level of scrutiny.
From Automation to Risk-Based Valuation
This is where the industry has an opportunity to move beyond the old automation-versus-human-review debate. The better approach isn’t to automate everything or manually review everything. It is to apply the right level of valuation and diligence to the decision and the risk.
For a straightforward property in a well-supported market with strong data and a valuation that performs as expected, an automated result may provide an efficient and appropriate answer. For another property, the right next step may be different. It could mean another valuation methodology, an evaluation or desktop valuation, additional property information, a virtual inspection or professional review depending on the decision and the level of risk.
Veros and Valligent have capabilities across that spectrum, from AVMs and portfolio analytics to evaluations, desktop valuations, inspections, appraisals and appraisal review. That breadth is important because no single valuation method is necessarily right for every property or every business decision. The opportunity is not to promote one approach over another, but to connect these capabilities intelligently so organizations can use the information that is most relevant to the decision they are making.
The same principle applies beyond origination. Servicers and investors may need to understand current values across large portfolios, identify areas of potential exposure and determine where additional analysis could be useful. Portfolio analytics can help surface properties or loans that meet defined risk criteria, while additional valuation or property-level information can provide greater context where it is needed.
The better question is not whether every property needs the deepest possible valuation. It is where additional information is most likely to change the decision.
The Real Advantage May Be Knowing When to Look Again
The mortgage industry’s investment in valuation technology has produced significant gains in speed, scale and access to information. The next opportunity may be less about simply producing valuations faster and more about becoming better at recognizing when the available information is sufficient and when it is not.
It requires a different way of thinking about automation. Technology can establish a strong baseline. Models can identify patterns. Property condition intelligence can add another dimension of context. Portfolio analytics can help identify areas that may warrant attention. Professional review or inspection can provide additional insight when the circumstances call for it. Experienced risk teams can then determine how much information is appropriate for the decision in front of them. That isn’t a failure of automation. It is the point of intelligent automation.
Closing the Valuation Trust Gap
The future of valuation isn’t about choosing between technology and expertise. It is about using technology to make better decisions about where expertise is most valuable.
The strongest valuation strategies will not necessarily be the ones that apply the same process to every property. They will be the ones that can move efficiently when the evidence supports moving efficiently, recognize when additional context is needed, and apply the appropriate level of diligence when it matters.
That is how the industry can close the Valuation Trust Gap: not by trusting technology less, but by understanding what the technology can tell us, what additional context can add, and when a closer look is warranted.
The goal isn’t more automation for its own sake. It is better decisions, supported by better data, stronger property context and the right level of judgment.
