Wednesday, February 28, 2024 | The Latest Buzz for the Appraisal Industry

Understanding Computer Vision: The Next Frontier in Property Valuation and Appraisal

Real estate’s valuation landscape is rapidly changing as two interconnected innovations are creating transformative waves: computer vision and artificial intelligence (AI). 

These game-changing technologies are bringing about remarkable improvements to the appraisal and valuation processes, enabling real estate professionals to interpret and utilize data in ways previously unimaginable. Computer vision is helping real estate appraisers see the unseen.

Computer vision empowered by AI is not just creating a ripple but has the potential, once unlocked, to generate a tsunami of change. While computer vision will not replace today’s real estate appraiser, appraisers who do not adopt computer vision will be replaced by those who do.





What is computer vision?

Computer vision is a subset of AI. It fundamentally changes how we interact with and interpret data extracted from visual sources. Computer vision is an AI technology that ‘teaches’ machines to understand images, similar to how humans do, by unlocking from images structured data points that can be used and analyzed.

The reason computer vision is being adopted so rapidly in our industry is because real estate is a visually driven industry. Every facet of the real estate ecosystem makes decisions by looking at listing photos, viewing virtual tours, zooming in on 3D models, and watching videos.

However, while it may only take a human a minute to review all the images for a property listing, it quickly becomes unscalable for a single person to analyze all the images of every property in an entire market. If you estimate that each listing in the US averages over two- dozen images per listing, even when listing inventory is historically low. October’s 737,000+ real estate listings translate into more than 18.4 million images uploaded.

Thus, computer vision’s ability to quickly and efficiently derive meaningful, actionable insights from millions of digital assets at scale represents a milestone breakthrough in the real estate industry.


How does it help a real estate professional?

AI, the broader umbrella under which computer vision falls, adds another layer of value. Artificial intelligence’s capacity for learning and adaptation means that these systems evolve: they become smarter, increase their efficiency, and are more effective the longer they are around and the more they are used.

In real estate, AI can generate predictive analytics that can forecast market trends, machine learning models that automate property valuations, natural language processing tools that generate compelling property descriptions, and image tagging, unlocking new data from the photos agents upload for their listings.

Today, computer vision and AI are already significantly impacting the real estate industry. Together, they are streamlining processes, providing more precise data, enhancing the customer experience, improving customer service, and opening new avenues for marketing and sales. From speeding up the listing process with auto-populated property details to providing a more friendly and personalized experience to potential buyers through immersive search, computer vision plus AI is reimagining the entire listing creation and promotion process.


Here are 6 valuation problems and challenges that computer vision can solve:

1.  Reduce loss mitigation: Inaccurate appraisals cost Appraisal Management Companies’ clients hundreds  of thousands of dollars, especially when inflated or undervalued appraisals are submitted for REO and Reverse Mortgages. Recently, we have witnessed Fannie Mae and Freddie Mac aggressively pushing back loans due to deficiencies in the appraisal. And now, Fannie and Freddie are focusing on the subject property’s condition and its comparables.

A computer vision solution is available today for AMCs to embed into their platforms to deliver nearly instant Fannie Mae-compliant image validation to appraisers. With built-in hard-stop rules that alert appraisers of potential discrepancies before submission, appraisal quality is improved, and loss mitigation is reduced.

2.  Increase client satisfaction: No one likes it when borrowers are inconvenienced more than once for an appraiser or data collector to enter their homes and take more photos. This new technology eliminates the possibility of leaving the property site with poor-quality pictures or incomplete data collection. This is an enormous time-saving benefit.

Avoiding making that second call to a homeowner can be a priceless benefit.

3.  Improve valuation accuracy: Inaccurate valuations of any kind can drastically increase the loss severity, particularly when a loan goes into default, is sold at auction, or endures another liquidation process. Improving the accuracy of the appraisal by better identifying the proper condition and quality of the property while reducing the time an appraiser has to spend collecting this data does three things: It  produces more accurate valuations, limits loss severity, and it dramatically reduces the time an appraiser would spend collecting the same data manually.

Computer vision can also be a trusted source for quality ratings; it can help automatically identify property damage, features not included in the report, and validate the information that is included, all improving valuation accuracy.

4.  Lower employee costs: Because computer vision creates a more efficient process, it saves time and lowers expenses. The right computer vision solution can pre-populate data through mobile apps to reduce the time spent by data collectors, lowering overall expenses. Additionally, rules engines can automatically review populated fields against details derived from photos with AI. Rather than requiring a manual review of an entire report, only the possible conflicts are highlighted which can be reviewed in a fraction of the time. 

Computer vision can significantly lower direct employee costs while increasing individual productivity when leveraged in this way.

5.  Leverages automation: AI and computer vision can do many mundane yet time-consuming things people do daily by leveraging automation. Creating efficient workflows to get the right level of complexity to the right person the first time dramatically reduces labor costs and review time. Through automating the review and decision-making processes, computer vision and AI reduce hours off each assignment, dramatically reducing turn times to deliver valuations. 

By automatically identifying and eliminating mistakes that would have triggered revisions in review, automation delivers a higher accuracy rate and lower per-order fulfillment costs for firms.

6.  Scalable solutions: Because technology works faster and more continuously than humans since it never takes a vacation or has a sick day, companies can expand their use of computer vision and AI rather than hiring, training, and growing their budget exponentially. When you realize that about 70% of an AMC’s budget is staffing, you can see how any double-digit reduction can immediately and positively impact the bottom line. With the recent changes in interest rates dramatically changing volumes, having a more scalable solution is essential to control costs. 

For appraisers, computer vision and AI is akin to having a virtual assistant instead of having to hire one to improve capacity and increase earnings.


The (near) future of computer vision

The question that every real estate professional needs to ask when adopting and adapting the use of any new technology is, “What problems, business challenges, or pain points can it solve?” The key is picking AI solutions that can help you overcome business obstacles.

Tom Armstrong, MAI

Is Commercial Work for You?

The Commercial Appraisal Path — What to Expect This article is for those interested in learning what it takes to add commercial and multifamily (4+

Read More »
Peter Christensen

An Inside Look at E&O

This article is featured in the 2023 edition of the Appraisal Buzz Magazine. Some other features we have in our magazine include funny Buzztoon comics,

Read More »