More and more of my clients arrive at our first conversation having already asked an AI.
They know what a notaire does. They have a rough price per square meter in mind for the 6th. They have read about the ten-day cooling-off period. Some have even run a building’s meeting minutes through a chatbot before we speak.
I understand why. I use these tools too. And the trend is clear: according to a Veterans United survey from June 2026, 45% of prospective American buyers have now used AI tools in their home search.
So the question is no longer whether my clients use AI. It is what AI can do for a Paris search, and where it stops.
What AI does remarkably well
Used well, AI is an excellent way to prepare.
- Understanding the process. The promesse de vente, the role of the notaire, the timeline from offer to keys: AI explains these clearly, at any hour, in plain English.
- Translating documents. Technical reports, building rules and meeting minutes are written in dense French. A first translation saves hours.
- Building a realistic budget. It can remind you to add notaire fees and transfer taxes, around 7.5% of the price, before you fall for an apartment you cannot afford.
- Comparing neighborhoods in broad strokes. The Marais, Saint-Germain, the 16th: AI can sketch the general character of each.
- Preparing your questions. This may be its most useful role. Arriving at a viewing with the right questions changes the conversation.
For all of this, I encourage clients to use it.
Contact me and let’s discuss what you’re looking for.
AI can describe Paris. It can't walk it.
Ask AI where to buy, and it will hand you the postcard: the Marais, Saint-Germain, the 7th, the Latin Quarter. Name any of them and it agrees enthusiastically, borrowing the same adjectives as every website: charming, vibrant, quintessentially Parisian.
Last spring, an American couple came to me beautifully prepared. AI had helped them define their search: a two-bedroom with outdoor space in the 6th, around €1.1 million. “Realistic,” it assured them.
On paper, yes. Averages per square metre are perfectly well-behaved. Paris apartments are not. Terraces in the 6th are rare, and rare things don’t sell at average prices. And their favourite listing left out one small detail: the building needed major façade works that had not yet been voted. Not in the photos, very much in the next owner’s bill.
So we started where a search should start: with the neighbourhood, not the postcard. A Sunday market or a Sunday hush? A café where the waiter remembers your order? A métro line that gets you to the opera in twelve minutes? The 7th is elegant and very quiet after 8 pm, which is perfect for some and a little too perfect for others.
We explored and compared arrondissements: hip, quiet, chic. They wound up in the northern 3rd, the end of the Marais the postcard never shows. The Marché des Enfants Rouges on a Saturday morning, galleries tucked into old courtyards, the Square du Temple as a neighbourhood garden rather than a tourist stop. It was not where they had pictured themselves, and it was much closer to what they wanted.
AI is a fine place to start. Finding where you’ll feel at home means walking the streets and experiencing the unique vibe of each quartier.
Most of what matters in a Paris search is never written online. It lives in conversations, documents and streets.
What AI cannot see
The Paris market still lives largely offline. That is where AI reaches its limits.
The apartments that never go online. In the most sought-after neighborhoods, a significant share of sales happen before a listing is ever published, through agents’ networks and direct introductions. An AI can only work with what has been written somewhere.
What the listing agent knows but will never write. The real margin for negotiation. Why the seller is selling, and how quickly. Whether there were truly “other offers.” This information exists only in conversations, and it is shared with people the agent trusts.
The document nobody has shared yet. AI can summarize the minutes of a co-owners’ meeting, but only once you have them. They often arrive late in the process. Knowing which ones to request, and what is missing from them, still takes experience.
The street at eleven at night. An AI can describe a neighborhood. It cannot walk you out of the building, turn right instead of left, and show you what the listing agent hoped you would not notice.
The price of this apartment, today. AI works from averages and past sales. But in Paris, a top floor with a terrace and a ground floor on the courtyard of the same building are two different markets. The right price is set by the specific apartment, the specific seller and the specific moment.
How to use AI well when buying in Paris
My advice is simple: use AI to prepare, and people to decide.
Ask it to explain the steps. Ask it to translate documents and suggest questions. Then check every answer that will influence a decision with someone who works in the market every day.
And if you have already done your research with AI, bring it to our first conversation. I am happy to go through it with you, line by line: what holds up, what does not, and what no tool could have told you.
AI can tell you how Paris works. It cannot tell you which door to knock on, or what is behind it. That part still happens on the ground.
Sources
Veterans United, Home Buyer Survey, June 2026.
Tell us about your property search
Whether you’re just starting or ready to buy, we’re here to help you navigate the process
