Whenever I travel, I love looking at homes, but on a recent trip to Paris, that was practically impossible. That’s because in Europe, listings live across a patchwork of brokerage sites and localized systems, with far less centralized access than the MLS-and-portal ecosystem we’re familiar with in the U.S.
Because of that experience, I started wondering how buyers in international markets ever find anything at all. My subsequent search led me to a major European real estate network that’s reframing how property discovery works by making its entire property catalog natively readable by AI systems like ChatGPT, Google AI Overview and Claude.
Not indexed or scraped. Structured for machine reasoning.
The tech deployment, built with agentic AI platform Kleio and rolled out across Orpi’s 1,250 real estate agencies in France, points to an emerging shift: Real estate search is moving away from portals and keyword filters toward conversational AI systems that interpret user intent directly.
If that model spreads, the front door to a real estate listing may no longer be your website, an IDX feed or a major portal. It may be a direct conversation with the AI platform the buyer prefers.
From portals to AI discovery
Orpi’s real estate catalog can now be accessed and interpreted by LLMs such as ChatGPT, Google AI, Claude and others. Instead of simply publishing listings on websites and portals, the system restructures them so AI agents can reason across both structured fields and narrative property descriptions.
Before being extended to external assistants like ChatGPT and Claude, Kleio’s AI was first deployed on Orpi’s own listing portal and agency websites. This allowed it to codify Orpi’s sales expertise into the AI, theoretically allowing it to promote the right properties and services to each customer.
Philippe Wellens
According to Philippe Wellens, co-founder and CEO at Kleio, the system developed for Orpi rebuilds how property data is stored and consumed, using agentic AI to analyze projects, advise users, and recommend properties, services and solutions. That allows Kleio’s Knowledge Engine to add function to the fragmented facts of property, pricing and document data associated with listings.
User intent replaces filters
Wellens framed the difference as intent-based discovery rather than traditional filter-based search. Since AI agents have trouble differentiating between structured and freeform information, Kleio’s combination of data and narrative enables conversational matching for buyers searching with AI tools.
While some portals in the U.S. have created standalone integrations with ChatGPT, Wellens said that Europe’s regulatory environment and market structure provided the opportunity for more flexibility in deploying Kleio’s service. “In Europe it is much less regulated than in the U.S. with the MLS and compliance requirements,” he said.
Without a centralized MLS, European networks had more control over their data and more urgency to reclaim distribution from online search portals. Brokerages in Europe “are bleeding money to these aggregators,” Wellens said, “so they are very incentivized to change that.”
AI search as a new distribution channel
As the buyer journey moves away from owned channels like brokerage websites and online portals, AI search becomes more important. AI search engines and agents are looking for property data that is unified, structured and accessible.
“A user formulating a real estate search directly with these assistants will be able to receive recommendations of Orpi properties in real time,” Orpi said in a media statement announcing the partnership with Kleio. That offers a visibility opportunity that traditional search engine ranking and paid portals can’t match.
“We are redistributing the game,” Wellens said, “because for the last 20 years [buyers] had to go through aggregators.” Now, he said, AI assistants themselves become the entry point.
“The significance of initiatives like Orpi and Kleio isn’t simply that property search is becoming AI-enabled,” Natasha Terinova of REACH UK and Second Century Ventures said. “It’s a response to the fact that the entire fabric of search is changing.”
“Real estate data is moving from static websites and documents into systems that can understand, interpret and act on information,” she said.
In the UK, Terinova said they’re seeing government-backed reforms focused on upfront information, digital property records and more efficient transactions. “Together these developments point toward a more connected and data-driven property ecosystem,” Terinova said.
If AI systems become the place where intent is formed and refined, then real estate discovery stops being a search problem and becomes a language problem. That means that whoever owns the most legible, structured and semantically rich property data may own the first impression.
Troy Palmquist is the founder and principal at HomeCode Advisors. Connect with him on LinkedIn.
