Guide
Send a customer who said "next to it works too" a listing across the city, and trust dies in a single message. What does area-aware matching actually rest on — and where does it make the difference?
Realtiq Team · notes from the field
Last updated · July 8, 2026
A customer writes: “I’m looking for a 2+1 in Bostanlı — right next to it works too.” For an ordinary matching system, the word “nearby” is an open invitation: every listing in the same city becomes a candidate, and the customer finds themselves looking at a property in a coastal town 100 kilometres away. One wrong suggestion changes the tone of the whole conversation — the customer now trusts themselves, not the list. What we call region intelligence exists to prevent exactly that break.
The root of the problem: when a customer says “nearby”, they mean physical closeness — walking distance, the adjacent neighbourhood, the same living area. Simple systems read “nearby” through administrative labels instead: same city, or failing that, same district. These are not the same thing. Two ends of one district can be forty minutes apart by car, while the border neighbourhoods of two different districts can sit five minutes’ walk from each other. Administrative labels don’t always reflect physical reality.
The Realtiq assistant knows the Türkiye’s official geography: 81 provinces, 973 districts, more than 73 thousand neighbourhoods and over a million streets. This knowledge rests not on guesswork but on a source aligned with the state’s address and statistics registries. When a customer types a neighbourhood name, the system knows which district and province it belongs to, where its boundaries run, and which neighbourhoods are genuinely adjacent to it. The AI is never allowed to “guess” geography; the official record always has the last word.
Every candidate listing lands in one of five tiers relative to the requested area. Ranking follows these tiers, and the most critical rule comes last: a listing in the “far” tier never makes the list at all. A customer asking for Bostanlı may be offered Atakent right next door — but the opposite shore of the bay, or a small town 100 kilometres out, is never presented as a “nearby alternative”.
A subtle detail: adjacency doesn’t respect district borders. For a customer, a listing just across the district line — but effectively within walking distance — ranks ahead of one in a remote corner of the same district. Because the customer isn’t shopping for a district label; they’re shopping for a living area.
Neighbourhood names repeat abundantly across Türkiye: Bahçelievler, Cumhuriyet, Yenimahalle… the same name exists in dozens of cities. When a customer asks “anything in Bahçelievler?”, the system pins the search to the city where your office’s portfolio actually is; İstanbul’s Bahçelievler never gets mixed up with İzmir’s. In the rare case where it genuinely cannot tell which city is meant, the system asks the customer instead of guessing. Saying “found it” in the wrong city costs more than finding nothing at all.
Nobody in the field writes in land-registry language. Turkish characters, dots, abbreviations, suffixes… The system tolerates these spelling variations and matches what the customer wrote to the official registry name. A “no results” answer is never caused by a spelling difference.
An exact match doesn’t always exist — and when it doesn’t, the worst option is to pretend otherwise and send irrelevant listings. Instead, the assistant states the situation plainly: “No exact match in Bostanlı right now; there are two listings in Atakent, right next door.” The alternative is clearly labelled as “adjacent”. The customer always knows what is an exact match and what is a nearby alternative — and that transparency builds far more trust than a few extra listings ever could.
The impact on the office fits in three sentences. Wasted showings drop: customers don’t look at homes in places they won’t go, and agents don’t book viewings that won’t happen. Conversations get shorter: the message rounds spent on “that area is too far for us” corrections disappear. And trust compounds: an assistant whose “nearby” turns out to be genuinely nearby becomes, in the customer’s eyes, an extension of your office. Region intelligence isn’t a showroom ornament — it’s the base layer that decides how accurate matching can be.
6 min read
A buyer who likes a listing almost always makes their first move with a WhatsApp message. What drives this behaviour, and how should it change your office's response flow?
Read the article →5 min read
A rigid budget filter drops the home a customer would fall in love with — over a mere 200 thousand lira. How does a 10% tolerance close this blind spot, and where should it stop?
Read the article →