Real estate AI software in Canada and the US
68% of real estate agents across North America have now used AI tools in some form, according to NAR's 2025 Technology Survey. Adoption of real estate AI software in Canada and the US is accelerating, and it is concentrated on drafting, scheduling, and keeping client relationships from going cold — not on replacing them.
Key points
- AI has moved past the experiment phase in brokerages across the US and Canada.
- The future of real estate Canada and the US are heading into is being shaped by two forces at once: new technology and new rules.
- The agents worth watching are not the ones chasing every new tool.
It always comes up near the end of the panel. Someone leans in and asks, "Should we actually be worried about AI?" The question is honest, and the answer is more useful than most of the hot takes pretending to have it solved.
What is AI already doing in real estate, and what is it not doing?
AI has moved past the experiment phase in brokerages across the US and Canada. According to the National Association of Realtors' 2025 Technology Survey, 68% of agents have now used AI tools in some form. Most of that use has come from small, practical uses: drafting listing descriptions, summarising long email threads, generating social posts, and keeping follow-ups from slipping.
That is the part of the job AI handles well. What it doesn't handle is the part where an agent sits at a kitchen table and reads the room before suggesting a number. Software can't tell when a seller is grieving a divorce rather than a price. It can't call a buyer and deliver the hard thing gently.
Agents who use these tools to keep their relationships in better order see value quickly. Those who expect software to run deals or replace years of trust walk away unimpressed. The line between those two outcomes has little to do with the tool and everything to do with how it is used.
What does the future of real estate look like across North America?
The future of real estate Canada and the US are heading into is being shaped by two forces at once: new technology and new rules. In the US, the post-NAR settlement world has reset how buyer-side compensation is discussed and pushed agents to articulate their value in plainer terms. Canadian agents face their own shifts, including tighter CREA rules on representation disclosure and provincial attention on AI-generated advertising.
Both markets are relationship-driven in a way that protects the agent's role. Home ownership rates, commission structures, mortgage rules, and tax treatment differ meaningfully across the border. Software still cannot explain a Quebec notary system to a first-time buyer from Ontario, or walk a Texas seller through a bidding-war counter at 9pm on a Tuesday.
What technology does change is how many relationships a single agent can hold properly. This is where AI adoption Canadian realtors are reporting matters most. Agents using it well are not closing dramatically differently. They are staying in touch with more people, more often, which is what turns a contact list into repeat business and referrals rather than a spreadsheet nobody opens.
Which agents are getting the most from AI right now?
The agents worth watching are not the ones chasing every new tool. They are the ones quietly sorting through which parts of their week no longer need their full attention, and putting the recovered time back into the people already in their contact list. Follow-ups, inbox triage, the past client who hasn't heard anything since the closing — this is the ground Worthington covers, drafting in the agent's voice for the agent to review and send.
A good tool here should feel boring, in the best sense. It should shorten the gap between what a client asks and when they hear back. It should remember the small commitments that make an agent trustworthy. And it should never be mistaken for the agent: the drafts go out from the agent's own account, and nobody on the client side is talking to software.
Chicago Agent Magazine put it well: AI will not replace real estate agents, but it will divide them. Agents who fold it into their weekly routine will look more capable each year. Those who ignore it will look slower, and slower is expensive in this business.
Questions agents ask about AI in North American real estate
Will AI replace real estate agents in the next five years?
No. Surveys from NAR, Delta Media, and Inman all point in the same direction: adoption is accelerating, but AI is being used to reduce administrative load rather than replace client relationships. The agents most at risk are not those competing with AI, but those competing with other agents who already use it well.
How fast is AI adoption among Canadian realtors?
Slower than the US, but catching up. Statistics Canada data shows the share of real estate and leasing firms planning to adopt AI nearly doubled, from just under 11% to 18.6%, between the 2024 and Q2 2025 periods. Canadian agents tend to adopt proven tools rather than early ones, so the curve is steeper now that the category has matured.
What tasks should a real estate agent actually use AI for?
Client follow-ups, calendar scheduling, drafting emails for review, summarising long threads, writing first-draft listing copy, and preparing for appointments. Anything repetitive and text-heavy is a reasonable starting point. Pricing decisions, negotiation, and client advice stay with the agent — and any tool claiming to set a listing price or produce a valuation is overstating what comparable-sales research can do.
Do clients mind when their agent uses AI?
Most do not, provided it is used to improve responsiveness rather than replace contact. What clients dislike is generic, obviously automated communication. A quick, human-sounding message drafted with help beats a slow, perfect one that arrives two days late.
What is the single biggest risk of ignoring AI right now?
Falling behind on response times, and losing touch with past clients. Buyers and sellers increasingly expect replies within minutes, not hours. Agents doing everything manually work longer weeks for the same outcome, and the competition is not a robot — it is another agent whose past clients hear from them four times a year.
Back to the question at the panel. The agents asking "should I be worried?" are asking the wrong one. A better question is which relationships are going quiet while you work live deals, and what it would take to keep them warm. That is the work Worthington does. If that sounds worth a look, worthington.ai is a sensible place to start.