The two real estate founders we hosted on Built This Week use AI for the slow, manual part of property work: research and inspection. Jackson Reiter of Realm AI runs dozens of agents that gather scattered data on commercial properties and cite every source. Zach Rattner of Yembo uses computer vision to turn a short phone video of a home into an inventory, a 3D model and floor plans. Both describe an industry that, until recently, ran on PDFs, printouts and manual typing.
This is a way for them to do instead of taking two months to actually diligence those properties, our platform allows them to do all 10 of them in a number of hours, sometimes even minutes per property.
— Jackson Reiter, Episode 47
How companies we talked to use AI in real estate
Realm AI: due diligence agents for commercial property
Jackson's starting point is that commercial real estate data is fragmented, with "no single source of truth." Realm's self-serve platform lets an analyst search any US property (Canada is coming) and get tax information, ownership records and a full ten-year pro forma, with every source cited. The same agents run inside Excel through a plug-in their CTO built. For a deal that moves forward, the Realm Report goes "10 times deeper" as a one-to-one service. (Episode 47, from 02:24)
At 06:59 Jackson described how the agents find data: a waterfall that starts with direct data-feed partners, then public sources and government sources such as New York City's online housing data. One thing Realm refuses to do is recommend deals. When Sam asked at 08:33 whether Realm could find properties that match a firm's criteria, Jackson said no, because those criteria are "the secret sauce" firms see as their edge.
Yembo: computer vision inside the home
Yembo sells to moving companies and property insurers. Instead of a 60 to 90-minute in-home visit, the homeowner gets a texted link and scans each room in the phone's browser. The AI identifies items and builds a "visual inventory" with numbered, color-coded photos, packing materials and items marked moving or not moving. For insurance claims, the same capture produces a 3D model and floor plans. The moving product runs in about 35 countries; insurance is US-only for now. (Episode 49, from 02:55)
The basic reason that we exist is we can compress down these tedious and complicated in home inspections that can take hours and cost billions, and we've shrunk it down into a process that takes just, ten, fifteen minutes.
— Zach Rattner, Episode 49
Jordan's take was that the trick is who touches the AI. The homeowner runs the scan, and the mover gets a better report without changing how they work or ever needing to know AI is involved.
What we built
We haven't built a real estate tool on the show yet. Both episodes were live product demos from the guests rather than host builds. For neighboring industries, see AI in insurance, where Yembo also appears, and AI in construction.
What's still hard
- The industry digitized but didn't automate. Jackson said firms still log into a tax assessor's portal, download a PDF and type values into Excel. Many like Claude for basic automation, but the work still lives in spreadsheets. He recalled a sophisticated New Jersey investor modeling a deal on a printed spreadsheet with a button calculator, less than five years earlier.
- Coverage gaps. Realm says it covers nearly everything relevant to a deal, but rural and underserved areas can have discrepancies. Personal data on owners is off-limits, and Jackson said they would never provide it.
- Big firms move slowly. Small firms find Realm and are running in a couple of weeks. Large firms adopt more slowly because of layers of complexity.
- First-time users. Yembo's users scan a home once, maybe once a decade. Zach said about 30% of Yembo's patents cover guiding the camera capture so the AI sees what it needs without a tutorial.
The upside both guests pointed to is doing work that used to be too expensive to do at all. Zach's example: a destination mover once emailed asking who had spent hours building a Yembo report by hand. Nobody had; the AI does it on every move.
Episodes on AI in real estate
- Episode 47: AI Agents Doing Commercial Real Estate Due Diligence in Minutes with Jackson Reiter of Realm AI (June 2026).
- Episode 49: A Ten Minute Phone Video Replaces Hours of Home Inspections with Zach Rattner of Yembo (July 2026).
FAQ
How is AI used in real estate?
The two founders we hosted use it for research and inspection. Realm AI's agents gather tax, ownership and financial data on any US commercial property and build a ten-year pro forma with sources cited; Yembo's computer vision turns a homeowner's phone video into an item inventory, 3D model and floor plans.
Can AI do commercial real estate due diligence?
Realm AI co-founder Jackson Reiter says an analyst with 10 deals on their desk can diligence them in hours instead of two months, sometimes minutes per property. For a deal that moves forward, Realm sells a deeper, white-glove Realm Report.
Will AI recommend which properties to buy?
Realm deliberately won't. Jackson said firms see their buying criteria as their competitive advantage, so Realm stays a research tool rather than a recommendation engine.
Who benefits most from AI in commercial real estate?
According to Jackson, small and medium investors: firms with two principals and three analysts can get to the deal table faster than big firms with 100 analysts, and they adopt in a couple of weeks while large firms move slower.