Yembo CTO Zach Rattner shows how a 10 to 15 minute phone video replaces hours-long in-home inspections for movers and property insurers using computer vision.
AI can turn a 10 to 15 minute phone video of a home into the inventory or 3D model that used to require an in-person inspection lasting hours. Zach Rattner, CTO and cofounder of Yembo, demoed exactly that on Episode 49 of Built This Week. Yembo sells its computer vision to moving companies and property insurers.
Zach Rattner says Yembo builds computer vision that deeply understands the interior of homes. In-home inspections are tedious and complicated, can take hours, and cost billions across the industry. Yembo compresses that into a short video scan. For a move, the output is an inventory list. For an insurance claim reconstruction estimate, it is a 3D model of the space and floor plans.
The moving product is deployed in about 35 countries and does thousands of inspections every day, concentrated in North America with presence in Europe, Asia and Oceania. Property insurance is more heavily regulated, so that product is US only for now, though Zach says the team is scoping expansion. You can find the company at yembo.ai.
Watch Zach Rattner demo Yembo's AI moving inventory in Episode 49 (starts at 2:55) · Episode 49 notes and transcript
Zach walked through a move from New York to London. Traditionally a company schedules a visit, someone rings your doorbell and writes down everything that needs to move. That list decides the truck size and how many cartons to bring, and it matters for customs on international moves and DOT regulations on domestic ones. The visit takes 60 to 90 minutes.
With Yembo, the customer gets a link by text and scans quick videos in their phone's web browser. The AI identifies what is in each room and builds the inventory with actual pictures, not a handwritten list. Zach says that helps the crew bring the right equipment, works as a checklist for loading and unloading, and gives a record to compare against if something is lost or damaged.
Zach shared his screen to show what Yembo calls the visual inventory. It summarizes the packing materials the AI calculated from the items in each room, splits items into moving and not moving, and shows film strips of images with every item color coded and numbered. Items marked as not moving are date and time stamped.
His favorite story is from around 2020. A customer was moving from Washington DC to London, and the mover at the destination emailed Yembo asking who on earth had the time to make the report. They assumed someone had spent hours taking pictures and adding the colored numbers by hand. Yembo told them it happens on every move, done by computer vision rather than a person.
That story led to Zach's main point about where AI creates value.
I think the obvious thing to do when people hear about AI is to automate and replace and just do the same old things with some percentage faster.
— Zach Rattner, Episode 49
The more compelling move, he says, is to use the fact that these things got easier and faster to do higher quality work, or work that was once cost prohibitive or labor intensive. Baking AI into the workflow lets you rethink how everything works.
Jordan called the model "pretty magical." Moving is an antiquated, paper-heavy industry, and handing AI to movers directly would probably fail because they don't use technology in their workflows. Yembo puts the tool in the homeowner's hands instead, so the mover gets a better document without ever needing to know AI exists. Our AI in logistics and AI in insurance pages cover other companies working in these industries.
Zach says he has Claude open from the moment he gets to his desk until he leaves, and sometimes on his phone after. Yembo's users are the opposite. Nobody deploys AI just because it is AI; they have quotas and work to get done. A homeowner scanning their house has never done it before and may not do it again for ten years.
I think probably 30% of my company's patents are on what do you do during the camera capture process to make sure that the AI sees what it needs to see, but without making somebody, like, watch a two hour tutorial video or read a manual beforehand.
— Zach Rattner, Episode 49
He thinks this is often overlooked. When AI is deployed into the physical world, there is usually no mature infrastructure around people using it, little awareness of what the technology does, and no clear limits on what it can and can't do. Jordan's summary: people know how to use a phone, but not how to prepare a home inventory for a professional international move.
In the news segment Sam noted that Fable 5 had been pulled and then authorized again, but as he understood it not for coding. Jordan used Fable 5 while it was available, loved it, and says the current models are not nearly as good. His read is that frontier labs are building their own framework with the government to avoid having models taken down, even without formal regulation in place. We tested Fable on the show in Episode 46.
Zach was not surprised by the pause.
In my view, I think the AI industry is sort of reaching adolescence. If you look at other markets like the automotive industry, it would sort of be unheard of to launch a new car, but then not have safety testing done or to launch a new drug in the pharmaceutical world, but then not exactly know the implications.
— Zach Rattner, Episode 49
He sympathizes with labs racing each other and China, and would like clearer guidance. Two or three years ago pulling a model meant losing an interesting toy. Now it means people lose the ability to get real work done, which is why he thinks AI is starting to look more like a public utility.
Jordan played devil's advocate. He compared AI labs to top cloud providers: you can't create a brand new AWS account and immediately launch or scale big servers without permissions and approvals. VC-backed labs chasing growth, he argues, didn't put the same know-your-customer controls, IP checks and customer vetting in place, citing news reports of foreign accounts abusing the model. With more controls earlier, he thinks this might not have happened.
He is not sure whether AI should be regulated like pharmaceuticals and cars, or treated like a coding language or database that shouldn't be. For him it is too early to tell. Sam's take was that the industry is entering its adolescence and a defined rollout framework for powerful models makes sense but doesn't really exist yet.
Zach's screen share of the visual inventory report is worth seeing on video. Watch Episode 49 of Built This Week, where the full transcript is also on the page, or read more about Zach Rattner.
