Every home listed for sale today comes with a visual record: room-by-room photos, condition details, sometimes a floor plan. That record is built for buyers, and the moment a purchase contract is signed, it is set aside. Kenon Chen, EVP of Strategy and Growth at Clear Capital, is working on what happens to that data next.
When a lender orders an appraisal, someone has to visit the property, take new photos, and manually assess condition and quality, work that in many cases already exists in the listing. The question the industry is starting to ask is whether machines can read that existing imagery well enough to support lending decisions, and whether lenders are ready to trust them.
The listing process has become remarkably visual. Buyers routinely tour homes through photo galleries before scheduling an in-person visit, and a listing without images feels incomplete. “When you’re browsing a listing on your favorite real estate portal, you’d be very surprised if there are no photos there,” Chen says. “It wouldn’t even make sense.”
Yet when a transaction moves from the marketing stage to the lending pipeline, that visual record no longer accompanies it. Lenders and appraisers start from scratch, dispatching inspectors, collecting new photos, and building independent assessments of the same property that was just photographed for the MLS. The result is a system in which the same ground is covered twice, at the buyer’s expense in both time and money.
The harder problem in property assessment isn’t identifying rooms. It’s evaluating them. Condition and quality are the variables that most directly affect value, and they have historically required human judgment to measure. Chen describes the traditional process as “humans staring and comparing photos to make sense of it,” which is slow, inconsistent, and difficult to scale.
Computer vision is changing that. Trained models can analyze a photo and extract not just the type of room it shows, but also indicators of its condition and finish quality. As Chen explains, condition and quality “are one of the most important factors that go into not only assessing the property you’re looking at, but also doing an accurate comparable analysis looking at comparable sales and listings that help you support an accurate value.” A flawed comparable can distort a valuation as much as a flawed inspection.
The more significant opportunity isn’t any single faster appraisal. It’s what happens when property data stops being recreated at every stage of a transaction. A home gets listed, sold, refinanced, and eventually sold again. Each event currently triggers its own independent data collection process, disconnected from what came before.
Chen frames the goal as “removing friction as it moves through that lifecycle and having a higher quality, higher fidelity understanding of the property early so there’s certainty up front, and you don’t have to keep doing the same duplicate steps downstream.” Floor plan data, through CubiCasa, acquired by Clear Capital in 2021, is following the same path. CubiCasa’s stated mission is to attach a floor plan to every residential listing, making spatial data a standard part of the listing record rather than an occasional add-on. Condition and quality scoring is the next layer. Clear Capital’s acquisition of Restb.ai further deepens this capability, adding computer vision technology purpose-built for property photo analysis and reinforcing the company’s broader investment in photo-driven property intelligence.
For borrowers, the effects of a slower, more redundant valuation process are concrete: longer closing timelines, more scheduling friction, and decisions that can hinge on a single inspector’s visit on a single day. Automating condition assessment and comparative analysis using existing data doesn’t just reduce costs for lenders. It compresses the window between offer and close and reduces the number of variables that can derail a transaction.
Chen keeps the end consumer in view. “Everything we’re doing downstream to speed up decision-making also has a positive impact on communities and on how people make decisions for real estate,” he says. Whether lenders gain enough confidence in automated assessments to act on them is still an open question, shaped in part by how secondary market participants such as Fannie Mae and Freddie Mac choose to treat photo-derived condition data in their underwriting guidelines. Millions of listing photos already exist. The remaining question is whether the industry will use them.
About the Expert:Kenon Chen is EVP of Strategy and Growth at Clear Capital, a real estate valuation and data analytics company.
This article is based on information provided by the expert source cited above. It is intended for general informational purposes only and does not constitute legal, financial, or real estate advice. Readers should conduct their own research and consult qualified professionals before making any real estate or financial decisions.
Explore similar articles from Our Team of Experts.
Even well-funded proptech companies with nearly $40 million in backing can encounter capital allocation challenges that force strategic pivots. Rentberry’s recent withdrawal from modular h...
Commercial real estate in smaller southern Illinois markets is experiencing rapid growth, driven by workforce availability, infrastructure, and demand for new technologies. BARBERMURPHY, a r...
