AI Data Broker for Real Estate: Listing Photos, Floor Plans & 3D Tours
Years of listing media, sitting in folders. Some of it may be licensable.
As an AI data broker for real estate, Signal & Form helps Canadian photography studios, media companies, and brokerages license listing imagery to AI developers. Real estate data licensing for AI is a new option for photography studios, media companies, and brokerages holding years of interior photos, floor plans, and 3D tours. AI developers use this kind of imagery to teach models how real homes look and how indoor spaces fit together. We help Canadian real estate businesses find out what they actually own, what it is worth, and how to license it safely.
You have thousands of listings on file. Can any of it earn again?
If you shoot real estate, run a media company, or manage a brokerage, you likely have an archive like this: tens of thousands of interior and exterior photos, floor plans, drone shots, and virtual tours. Each listing sold, and the media went into a folder.
Now you see AI tools generating home interiors and staging rooms, and you wonder whether your work trained them, and whether you could get paid for the archive instead. Those are fair questions.
The honest answer starts with ownership. In real estate, the person who paid for the photos is often not the person who owns them, and listing content shared through MLS systems comes with its own rules. Sorting that out is the first job.
What Sells
Data in real estate that AI buyers want
Interior and exterior photo libraries
Large, consistently shot sets of real rooms, kitchens, baths, and exteriors across many property types. Image libraries are licensed to AI companies at scale; Shutterstock's early deals with Apple, Meta, Google, and Amazon were reported at US$25 million to $50 million each.
Floor plans
Measured plans, ideally with rooms and features labelled. Researchers built annotated floor plan datasets such as CubiCasa5K, with 5,000 plans tagged into more than 80 object categories, to train models that read building layouts.
3D scans and virtual tours
Walkthrough captures of real interiors. The Habitat-Matterport 3D dataset of 1,000 building-scale reconstructions, including homes, was built for embodied AI: training agents and robots to navigate indoor spaces.
Drone and aerial imagery
Neighbourhood, lot, and roof views. Aerial footage is among the material reported to earn premium rates when AI companies buy video.
Who wants real estate imagery, and why
Three kinds of buyers are relevant. Companies building image and video models want large, varied sets of real interiors and exteriors. Developers of design, staging, and property tools want floor plans and room imagery that teach models layouts and finishes. And robotics and embodied AI teams want 3D captures of real homes, because a home robot has to learn to move through spaces it has never seen.
Real estate data is not one of the publicly reported mega deals, so be wary of anyone quoting a price per listing. What drives value is the same as in other categories: volume, consistency, metadata such as room labels and dimensions, and above all clean rights. A studio with 100,000 photos it owns outright is in a far stronger position than a brokerage with a million images it cannot prove it can license.
Before You Sell
Rights, MLS, and privacy issues specific to real estate
- Photographers usually own the copyright. The Canadian Real Estate Association notes that since a 2012 change to the Copyright Act, a photographer hired to shoot listings likely owns the copyright, not the brokerage that paid. A brokerage can only license images it holds through a written assignment or a licence broad enough to cover this use.
- MLS and DDF content cannot be used for AI training. CREA's rules for its Data Distribution Facility state that DDF listing content cannot be used to train an AI system, and that AI providers may use it only to deliver the specific feature a participant offers. Local board and MLS rules also apply. Data pulled from listing feeds is not yours to license.
- Listing descriptions. Remarks are written by agents under brokerage and board rules, and often published through MLS systems. Check who owns the text and what the board allows before treating it as licensable.
- Sellers and occupants. Interior photos can show family pictures, documents, artwork, faces, and other personal details. Under PIPEDA and BC PIPA, using personal information for a new purpose generally needs fresh consent, and Canada's privacy commissioners caution that publicly available data is not automatically free to use. Personal details should be removed or blurred, and addresses separated from images where possible.
- Quebec. Law 25 requires a privacy impact assessment before personal information is communicated outside Quebec, which matters if a Quebec dataset goes to a foreign buyer.
General information, not legal advice. Have a lawyer review rights, consent, and any licensing contract.
The Canadian picture
Canada has an unusually clear rule on listing content. CREA's DDF rules allow participants to offer some AI features, but they prohibit using DDF listing content to train AI and do not create any right to download or republish it. Anyone promising to sell "MLS data" to AI companies should be treated with caution.
That points the opportunity at the people who own their media. Because of the 2012 copyright change, in many cases that is the real estate photographer or media company rather than the brokerage. Studios that have shot thousands of Canadian homes, and brokerages with written assignments from their photographers, are the natural candidates for this kind of data monetization.
One more Canadian wrinkle: if a collection includes Quebec properties and personal information has not been fully removed, Law 25 requires a privacy impact assessment before that data goes to a buyer outside the province. Clean, de-identified imagery avoids most of that friction and is easier to license.
How It Works
From archive to licensing deal
We review what you hold, estimate a realistic value, work with your lawyer on rights and consent, document the dataset, and route it to the right buyers. You pay a success fee only if a deal closes. See the full process on our AI data licensing page.
FAQ
Questions about AI data broker for real estate
Only the photos it actually owns or holds a broad enough licence for. Since 2012, a photographer hired to shoot listings likely owns the copyright, according to CREA. If your brokerage has written assignments from its photographers, you may have a licensable archive. If not, the photographer is usually the one who can license it.
No. CREA's DDF rules state that DDF listing content cannot be used to train an AI system, and local board rules add their own limits. Listing feeds are licensed to you for specific uses, not owned by you. Real estate data licensing for AI has to start from media and data you created or acquired the rights to.
Often, yes, unless your contracts assigned the copyright to the client. Many studios license images to agents for marketing only and keep ownership. Check your terms with each client. If you kept the copyright, you may be able to license the archive, subject to privacy clean-up. This is general information, not legal advice.
The photos themselves are about property, but interiors often capture personal information such as family photos, documents, and occasionally people. Those details should be removed or blurred, and addresses separated from images where possible. Where personal information remains, privacy law generally requires consent for a new use. Your lawyer should confirm the approach.
They can be, because they capture how a space fits together, which is what robotics and spatial AI teams need. Research datasets of 3D home scans and annotated floor plans exist precisely because this data is hard to collect. Value still depends on volume, quality, labelling, and rights, and there is no public price list for it yet.
Related Reading
What AI Companies Pay for Training Data in 2026: Every Public Price, Sourced
How to Sell Your Data to AI Companies from Canada: Process, Buyers, and the Law
Egocentric Data for Robotics: What It Is, Who Buys It, and What It Pays
Other Industries
Sources: CREA Café: Re-using listing photographs (photographer copyright since 2012) · REALTOR.ca DDF: changes to terms and conditions (AI and augmentation rules) · Copyright Act, s. 13: ownership and assignment (Justice Laws) · Reuters via The Decoder: Shutterstock and other AI image licensing deals · Bloomberg via The Decoder: premium rates for drone and 4K footage (Jan 2025) · arXiv: CubiCasa5K floor plan dataset · arXiv: Habitat-Matterport 3D dataset for embodied AI · OPC, OIPC BC and OIPC Alberta: Guidelines for obtaining meaningful consent · OPC: Principles for responsible, trustworthy and privacy-protective generative AI · Quebec private sector privacy act, s. 17 (LegisQuébec)
Sitting on real estate data? Find out what it is worth.
Book a free call. Tell us what you hold, and we will give you an honest read on whether it is worth licensing.