Clear Capital is a mortgage valuation and analytics company that recently acquired Restb.ai, which uses computer vision to turn property images into structured data for real estate valuation and lending. The FR recently sat down with Kenon Chen, EVP of strategy and growth at Clear Capital, to discuss how the technology fits into the company’s broader strategy.

Clear Capital now owns CubiCasa, which creates digital floor plans, and Restb.ai, which uses AI to analyze property photos. How will these tools work with Clear Capital’s property data and valuation products, and what are you ultimately trying to build?

The housing industry is going through one of its largest transformations in decades. At the heart of this transformation is the shift towards a more digital-first mentality. The industry as a whole is learning to turn visual property data into trusted, decision-ready intelligence.

Clear Capital and Restb.ai joined forces to give real estate professionals a more complete and consistent picture of a property, helping them make better-informed decisions.

The vision is to reduce the “blind spots” that have historically made property decisions difficult. That means capturing property data through digital floor plans, turning photos into structured intelligence and applying valuation technology and analytics. Real estate professionals should be equipped with the data required for better decisions, smoother workflows, and more accurate alignment between buyers, lenders, and properties.

By integrating these, we can deliver a more connected suite of products that offer a complete picture of a property to the broader ecosystem.

Computer vision has been a promising technology in real estate for more than a decade, yet adoption has often lagged behind the hype. What has changed in AI model performance, data quality, regulatory acceptance and customer demand to make this the right time for Clear Capital to put computer vision at the core of its strategy?

The decision to modernize and more deeply incorporate computer vision is driven by an accelerating market demand for faster, more data-driven decisions in mortgage and real estate.

Computer vision has long held enormous potential for real estate, but until recently the industry and even the technology had not fully matured. While we’ve viewed the promise of computer vision technology, the timing reflects the industry's shift toward digital, data-driven processes and the need for a scalable, systematic way to extract signals from property imagery that were previously hard to analyze.

The market is evolving. Over the past few years, advances in foundation models, image recognition accuracy and multimodal AI have allowed computer vision to interpret property characteristics with a consistency and scale that wasn’t previously possible.

The industry is under increasing pressure to improve efficiency, reduce costs, and deliver faster decisions. Regulatory modernization efforts, including version 3.6 of the Uniform Appraisal Dataset, the standardized format for submitting appraisal data to Fannie Mae and Freddie Mac, are also encouraging greater standardization and digital-first workflows. This makes it easier to integrate AI-driven property insights into valuation processes.

Traditional automated valuation models, or AVMs, estimate a property’s value using structured property and transaction data. How does adding visual intelligence change the valuation equation, and where are you seeing the greatest gains in accuracy, confidence or risk detection?

Traditional AVMs have always been limited by what structured data can describe. While they capture a property's size, age, location, and sales history, AVMs will often miss its condition, quality, maintenance, and unique features that significantly influence value.

Property photos have valuable information that has historically been difficult to measure consistently. By integrating visual intelligence into the valuation process, those traditional datasets are enriched with objective property insights, resulting in a more complete and consistent view of every asset.

Combining structured data and advanced analytics with AI-powered visual signals can strengthen quality control and identify potential risks earlier while making valuations more explainable. Rather than replacing human expertise, AI gives appraisers and valuation professionals richer, more reliable inputs to complement traditional valuation models. Ultimately, it gives lenders, appraisers and investors greater confidence in the data behind their decisions and improves consistency and efficiency throughout the valuation process.

With CubiCasa capturing property layouts and Restb.ai extracting insights from imagery, you're collecting a much richer property dataset at the point of inspection. How does that change the lender workflow, and what manual steps or third-party processes do you believe can ultimately be eliminated?

Using technology that’s embedding and creating high quality data earlier in the process is transforming workflows. CubiCasa provides the real estate community and home buyers with a more accurate understanding of the property during the listing process and Restb.ai removes significant manual effort traditionally required to interpret imagery.

Over time, this will streamline communication among appraisers, lenders and other stakeholders by creating a shared, consistent understanding of the asset. The result is a more efficient workflow with shorter turn times and more consistent data, allowing valuation professionals to focus their expertise on analysis rather than data collection while strengthening risk management.

Looking ahead 12 to 18 months, where should customers expect to see the first meaningful impact from the acquisition? Are there specific valuation, underwriting, quality-control, or appraisal workflows where visual intelligence will be embedded first, and what business outcomes are you targeting?

The business outcomes we are targeting are clear: greater consistency in valuations, reduced operational friction, improved risk visibility, and faster, more confident lending decisions. By making visual data usable at scale, we are helping transform property imagery from a passive record into a powerful source of intelligence.

Subscribe to the FR

The first meaningful impact will come from bringing computer vision intelligence directly into the workflows where understanding a property's physical characteristics matters most. Turning images into structured, actionable insights allows them to be integrated into existing decisioning processes.

The goal is not to replace existing workflows, but to make them smarter and more efficient. Over time, we expect this combination to reduce manual review, accelerate decision cycles, improve data quality, and help lenders manage risk with greater transparency and confidence.

Restb.ai CEO and co-founder Xavi Hernando contributed insights.