Study · AI Visibility

Which real estate agent does AI recommend, district by district?

We asked ChatGPT, Perplexity, Gemini and Google the same question across six Vienna districts. The result: six districts, six completely different lists. The names that are famous city-wide show up in not a single district, and in most districts exactly one firm wins across all four AIs.

🇩🇪 Diese Studie auf Deutsch lesen

More and more people no longer ask Google, they ask AI: “Which real estate agent in Döbling is good?” Whoever gets named gets the call. We wanted to know whether there is one single answer, or whether the recommendation changes district by district. So we queried four AIs across six Vienna districts.

The result in one sentence

Six districts, the same question, six completely different agent lists. The names that are famous city-wide appear in not a single district. And where one firm wins across all four AIs, it always has the same thing: lots of reviews and clean directory listings.

The experiment

On 3 July 2026, from a Vienna-based browser, we queried four AI engines in each of the six selected districts with the exact same wording, only the district was swapped:

The exact question to each AI, per district
“Which real estate agent in [district] Vienna would you recommend?”
Engines
ChatGPT, Perplexity, Gemini, Google (Local Pack)
Date
3 July 2026
Location
Vienna, Austria (real IP, no VPN, free tier)
Districts
1010 Innere Stadt, 1050 Margareten, 1130 Hietzing, 1190 Döbling, 1210 Floridsdorf, 1230 Liesing
Counted
Every real estate firm named, per district and engine

Finding 1: Every district is its own game

The clearest result first: there is no “best agent in Vienna” that AI names consistently. The moment you put the district into the question, the answer changes completely. No agent recommended in one district reappears in the next. Here is a selection of the firms the AIs name per district:

DistrictNamed by the four AIs (selection)
1010 Innere StadtOTTO Immobilien, ÖRAG, IMMOTOTAL, Stanzel & Co, Real Immo Wien, VON POLL. Perplexity names no specific agent here at all.
1050 Margareten4M Immobilien, OPTIN, Reinberg & Partner, LAGUN Realitäten, Feeling Home
1130 HietzingKlein & Partner, Real Vision, Wohncloud, IMMOTOTAL, REMAX First, CONCREET
1190 DöblingBRI Immobilienmakler, Kaltenegger Real, McMakler, Wolke 7, RT Immobilien, GOLDAS
1210 FloridsdorfRE/MAX Dreams, Perfect Deal, SMK Immo, Carma & Partner, Rustler, Bittendorfer
1230 LiesingSteinhoff, Zellmann, Reischel, Pluskota, EV Immobilien
Which agent the AI names for you depends almost entirely on which district you ask about. “Vienna” on its own does not exist for the AI.

Finding 2: In most districts, exactly one firm wins

This is where it gets interesting for agents. In three of the six districts, one firm was named by all four AIs. In two others there was a clear leader with three out of four. And in one district, the dense first, not a single firm broke through.

1050 Margareten4M Immobilien
4/4
1190 DöblingBRI Immobilienmakler
4/4
1230 LiesingSteinhoff / Reischel
4/4
1130 HietzingWohncloud
3/4
1210 FloridsdorfRE/MAX Dreams
3/4
1010 Innere Stadtno consensus
0/4
How many of the 4 AI engines agree on the same top firm per district. In the densest district (1010) there was no shared recommendation, while the outer districts had clear winners.

And here is what really matters: the firms that win everywhere all share the same profile. 4M Immobilien in Margareten has over 1,200 Google reviews at 4.9 stars. BRI in Döbling sits right in the 19th district with 4.9 stars and is listed on WKO, Herold and ImmoScout. Steinhoff and Reischel in Liesing, same picture. Always both together: lots of fresh reviews and clean directory listings. Whoever has only one of the two makes it into one or two AIs, but not all four.

In the 1st district, where competition is densest, Perplexity would not even commit and gave general tips instead of names. That fits the pattern: the denser the market, the harder a clear recommendation. For agents in the outer districts, that is good news.

Finding 3: Every AI has a mind of its own

Why six different worlds? Because each AI reads from a different source. Döbling shows it most clearly. Four AIs, the same question, and only one single firm (BRI) makes it into all four lists:

EngineNames in Döbling, among othersDraws on
ChatGPTRT Immobilien, BRI, MAITZ, GOLDAS, RE/MAX TodayGoogle Maps, dibeo.at, district portals
PerplexityBRI, Alexander Ringsmuth, McMakler, NEUWIRTH, Klemens BraunischImmoScout24, WKO, Herold
GeminiWolke 7, Marschall, Die IMMOREI, BRIGoogle Maps (map pins + stars)
GoogleKaltenegger, McMakler, BRI, Ringsmuth, GOLDAS, RT ImmobilienGoogle Local Pack (reviews)

Across all six districts, a clear pattern emerges of which AI reads from what:

  • ChatGPT and Google mainly read Google Maps and reviews. Their lists resemble each other closely and favor agents with lots of stars.
  • Perplexity reads almost exclusively business directories (ImmoScout24, WKO, Herold, FirmenABC) and therefore often names completely different firms than the three map-based AIs.
  • Gemini shows Google Maps pins with stars and, in prime locations, leans toward award-winning firms (OTTO, IMMOTOTAL, Klein & Partner with the IMMY Award).

The logic is simple: no source, no mention. A firm that is missing from an engine's sources will not be recommended by that engine, no matter how good it is. This matches what larger studies show: ChatGPT recommends only around 1 in 100 local businesses at all1, and the engines pull their recommendations from almost entirely different sources2.

What this means for Vienna's real estate agents

  • Visibility is local, not company-wide. You can lead across all four AIs in one district and be completely absent in the neighboring one. If you operate across areas, you have to be visible per target area.
  • There is a recipe. The firms that win everywhere have lots of fresh Google reviews and clean, consistent listings in the directories the AIs read from. Both together are the price of entry.
  • Every engine needs its own approach. A strong Google profile gets you into ChatGPT, Google and Gemini. For Perplexity you need the directories. A single channel is not enough.

That is exactly what Generative Engine Optimization (GEO) is: making sure you get named in every relevant AI and in your target district. That is what KIGEO specializes in.

Does AI recommend you in your district?

We check for free whether ChatGPT, Perplexity, Gemini & Google name you in your target area, and show which competitors get recommended instead.

free AI visibility check →

Methodology & sources

All four engines were queried on 3 July 2026 from a Vienna-based browser with the exact wording quoted above, one answer per district and engine, on the free tier. For Google, which returns no AI overview for this question, the local map and results block (Local Pack) was counted. Every real estate firm named was recorded. This is a sample, not a full census. The sources the engines drew on were essentially Google Maps as well as the directories ImmoScout24, WKO Firmen A-Z, Herold and FirmenABC, supplemented by district portals such as dibeo.at and BezirksTipp.at. Studies used for context:

  1. SOCi, Local Visibility Index 2026, context: across hundreds of thousands of locations, ChatGPT names only around 1.2 percent of local businesses in a recommendation (US dataset, indicative).
  2. NeuRank, AI Visibility Study 2026, context: ChatGPT bases around 38 percent of its local recommendations on Google Maps, while Perplexity cites almost exclusively German-language directories (German market).
  3. Engines and directories used for data collection: ChatGPT, Perplexity, Gemini, Google Local Pack; ImmoScout24, firmen.wko.at, Herold, FirmenABC, makler-empfehlung.at, willhaben, dibeo.at, BezirksTipp.at.

Download this study's raw data as CSV ↓

K

About KIGEO

KIGEO is an AI visibility (GEO) and SEO agency from Vienna. We make sure businesses get named in ChatGPT, Perplexity, Gemini and Google, in the right area and with correct details, and we measure it honestly. This study was produced with the same methods we use for our clients. Request a free check →