AI for real estate agents in Ontario: what RECO's new guidance allows
Ontario's real estate regulator has now said how agents and brokerages should use artificial intelligence. Real Estate Magazine reported on November 2, 2026 that RECO released the guidance in October (Real Estate Magazine). It doesn't ban anything. It says the existing rules apply to whatever AI produces, and that the agent is still the one responsible.
This guide covers what the guidance says, what that means for everyday marketing, and how AI search is changing which pages on a real estate website are worth having. The guidance was read on RECO's website on November 10, 2026.
What RECO's guidance says
Six points, in RECO's own words where quoted:
- You're still responsible. "The use of AI does not transfer professional responsibility to a technology provider." Brokerages and agents "remain accountable for all data, representations, and communications generated or assisted by AI tools."
- Check everything. "AI outputs may look polished but can contain content errors, outdated market information, or false data."
- Keep client information out of public tools. "Do not enter personal, financial, or confidential client/transaction data into public AI platforms." Before entering any personal details or contract terms into an AI platform, get the client's explicit, written consent.
- Advertising rules still apply. AI-generated posts, descriptions and enhanced visuals "must comply with all TRESA advertising obligations including the mandatory placement of brokerage identification."
- Label altered images. Virtual staging and AI modifications must "accurately reflect permanent features", and any digitally altered image must be clearly labelled.
- Get approval. Check that a tool "aligns with company policy and has Broker of Record approval". Brokerages "should establish clear internal policies, training, and an approved tool list".
The guidance cites the Trust in Real Estate Services Act and its Code of Ethics regulation.
What that means day to day
| Task | Fine to use AI? | What to do |
|---|---|---|
| Drafting a listing description | Yes | Check every fact against the listing: room count, lot size, upgrades, taxes |
| Drafting a market update or newsletter | Yes | Use published figures, name the source, and check each number yourself |
| Social posts and captions | Yes | Include the brokerage name, as in any advertising |
| Virtual staging | Yes | Keep permanent features accurate and label the image as altered |
| Removing a hydro pole or changing the view | No | That misrepresents a permanent feature |
| Summarising an offer or a client's finances in a public chatbot | No, without written consent | Keep personal and transaction details out of public tools |
| Answering a legal or tax question for a client | No | AI is for support and doesn't replace professional judgment |
Before using a new tool for any of this, ask whether your brokerage has approved it.
Where AI saves a team real time
The safest uses are the ones where you supply the facts and AI supplies the first draft:
- Turning published statistics into plain language. Paste in your board's monthly release and ask for a 100-word summary, then check it against the release. For October 2026, TRREB reported 5,603 GTA sales, down 7.9% from a year earlier, and new listings down 11.9% (TRREB, November 4, 2026).
- Rewriting your own notes. Your observations about a street or a building, tidied into a paragraph for a neighbourhood page.
- Reformatting. One article turned into a newsletter item, a LinkedIn post and three captions.
In each case the knowledge is yours and the accuracy is checkable. The risky uses are the opposite: asking AI for facts about a property, a market or a law, and publishing the answer unchecked.
AI search and your website
AI is also changing how buyers and sellers find information. Google now answers some searches with an AI summary above the results, so the searcher may never visit a website.
Real estate has been less affected than most industries so far. One vendor's benchmark found AI Overviews on about 4.5% of the real estate searches it tracked (Luxury Presence). The effect falls unevenly, though. General questions, like "is Leslieville a good place to live", are the kind an AI summary can answer without a click. Specific, current, local data is much harder to summarise away.
That has two consequences for a team's website:
- Generic area guides are worth less. If a page says what any summary could say, fewer people will need to visit it.
- Data you're licensed to show is worth more. Sold prices on a VOW sit behind a registration the board requires. An AI summary can't display them, so the person who wants to know what the house on their street sold for still has to visit a site and sign up. IDX, VOW and CREA DDF explained covers what each feed allows.
The pages that hold up are the ones built on things only you have: licensed sold data, your own market figures and first-hand knowledge of the streets. SEO for realtors in Ontario covers how those pages are built.
AI inside your CRM
Many CRMs now draft replies and summarise calls. Follow Up Boss, for example, says its AI features carry no extra charge but rely on its calling product. The same guidance applies: the agent is accountable for what's sent, and the tool needs the brokerage's approval. Our real estate CRM guide for Canadian teams compares prices and what to ask each vendor.
A short checklist
- Ask your broker of record which AI tools are approved.
- Keep client names, finances and contract terms out of public AI tools unless you have written consent.
- Check every AI draft against a source before it's published.
- Put the brokerage name on everything, AI-written or not.
- Label virtually staged and digitally altered images, and never alter a permanent feature.
- Build your site around data and knowledge an AI summary can't supply.
A neighbourhood portal is one way to do the last one: a site for one farm, built on sold data visitors register to see.
This is general information, not legal advice; check with your brokerage.
