TLDRocket
Sign in

How Scout24 is building the next generation of real-estate search with AI

OpenAI

Scout24 built an AI chat assistant using GPT-5 to help people search for homes and apartments. Instead of filters and dropdowns, it just asks you questions and finds what fits.

Based on reporting by OpenAI — read the original for the full story.

Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error

Anyone who has used a property portal knows the drill: pick a city, set a price ceiling, tick boxes for bedrooms and square meters, then wade through a hundred listings that almost match what you wanted. Scout24, the German company behind ImmoScout24, decided that process is broken enough to warrant rebuilding it around a language model rather than a filter panel.

The new assistant runs on GPT-5 and behaves less like a search engine and more like a knowledgeable friend who happens to know the local housing market cold. Tell it you need a two-bedroom place near good schools with a balcony, and it doesn't just parse keywords — it asks follow-up questions when your request is vague, the same way a real estate agent would probe for what you actually care about before showing you anything. It summarizes listings instead of dumping raw data, and it can explain tradeoffs, like why a cheaper apartment might mean a longer commute.

What's notable is the ambition behind the framing. Scout24 isn't pitching this as a chatbot bolted onto an existing site for novelty's sake. The company is describing it as a rethink of the entire search experience, one where natural conversation replaces the rigid, checkbox-driven interfaces that have defined property search since the earliest listing sites went online in the late 1990s. That's a big claim for a market that has resisted major UX changes for decades, largely because the underlying data — square footage, price, location — is inherently structured and filterable.

The bet Scout24 is making is that structured data isn't the same as structured need. People don't think in filters; they think in scenarios — a growing family, a shorter commute, a specific neighborhood vibe. A model like GPT-5 that can hold context across a conversation and reason about tradeoffs is arguably a better match for that kind of fuzzy, human decision-making than any dropdown menu ever was.

My take — AI-written commentary, not fact-checked reporting

This is the kind of AI deployment I actually want to see more of — not a flashy demo, but a genuine rethink of a clunky, decades-old workflow that real people deal with constantly. Real estate search has been begging for this since forever, and if GPT-5 can cut through the tedium of filter-hell, good. My only worry is whether Scout24 keeps the underlying listing data transparent and unbiased, or whether the 'helpful assistant' quietly nudges you toward paid placements — that's the tension every AI-powered marketplace will eventually have to answer for.

Read more about this at: OpenAI

Related stories

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.