AI you can
trust your clients with.
We're a Québec company building AI for one of the most personal decisions people make — their home. So we'd rather move carefully and openlythan fast and quiet. Here's how we keep your teammate honest, safe, and yours.
Six commitments,
in plain language.
No 40-page policy nobody reads. These are the rules our teammate lives by — and the ones you can hold us to.
A human is always in the loop
The teammate drafts, suggests and prepares. You approve, send and decide. It never speaks to a client or signs anything on its own.
Your data stays yours
Client information is encrypted, stored in Canada, and never used to train shared models. It works for you and no one else.
Transparent by design
It tells clients when they're talking to AI, and shows you where every number comes from — Centris comps, sources, the lot.
Fair to everyone
We test for bias against protected groups so the teammate never nudges a fair-housing line. Equal service is non-negotiable.
Privacy first, by law
Built to PIPEDA and Québec's Law 25 from day one — consent, access and deletion are features, not afterthoughts.
We own our mistakes
A real team in Québec, reachable by a real person. If the AI gets it wrong, we tell you and we fix it — no black box.
The frameworks
we're signing.
Anyone can say they're responsible. We'd rather be measured against standards written by people far smarter than us — Canadian ones, first.
Montréal Declaration for a Responsible Development of AI
A made-in-Québec charter for AI that serves well-being, autonomy and democratic participation. It's the standard our own city wrote — so it's the one we start with.
Voluntary Code of Conduct on Advanced Generative AI
Canada's federal code for the safe, transparent and accountable use of generative AI — covering safety testing, human oversight and clear labelling of AI content.
What happens to your
clients' information.
Short version: it stays in Canada, it stays encrypted, and it stays yours. We don't sell it, we don't mine it, and we never fold it into a shared model that anyone else could benefit from.
You can export or permanently delete everything we hold for you, any time, from your settings — and we'll confirm it's gone.
A little AI literacy,
since we're here.
We can't teach the whole field — but we can clear up the four questions agents actually ask us. Plain answers, no hype.
Will it replace my judgment?
No. Think of it as a sharp junior teammate: it gathers comps, drafts the email and remembers every detail — then hands it to you to decide. The relationship and the call are always yours.
What's an AI "hallucination"?
It's when AI states something confidently that isn't true. We reduce the risk by grounding answers in real Centris data and showing sources — and by keeping you in the loop to catch anything that looks off.
Can AI be biased?
Yes — AI learns from data, and data carries human bias. That's why we actively test for it, especially around protected groups and fair-housing rules, and why a human signs off on client-facing work.
Where does it get its information?
From the sources you'd use yourself — Centris listings, public market data and your own notes— not from scraping the open web. If it can't cite where a figure came from, it won't state it as fact.
Doing this right
is a conversation.
If something here matters to you — or you think we've missed something — we want to hear it. That's the whole point.