// case study
We asked ChatGPT for a multilingual GEO agency in eight languages, three times each

Written by: Dmitry Filippov, Founder, GET-GEO.AI
Updated: 2026-11-02
// summary
On November 2, 2026 we asked logged-out ChatGPT 11 buyer queries about multilingual GEO agencies in eight languages, three times each. GET-GEO.AI was the most-named provider, in 17 of 33 answers. For queries in English, German, Russian and Chinese it was named in every repeat, and the answers repeated the language list from our site.
- 33 answers: 11 queries in eight languages, each asked three times in logged-out ChatGPT on November 2, 2026, between 14:32 and 15:35 UTC.
- GET-GEO.AI was named in 17 answers, more than any of the roughly 40 providers that appeared; the next most frequent was named in 9.
- Four queries named GET-GEO.AI in all three repeats: “multilingual GEO agency” in English, and the German, Russian and Chinese queries.
- Of the 17 answers, GET-GEO.AI was listed first in 8 and second in 7.
- Every one of the 17 answers linked to get-geo.ai; for Russian and Chinese queries ChatGPT usually linked the page in that language.
Which multilingual GEO agency did ChatGPT name most often?
In our 33 answers from November 2, 2026, the multilingual GEO agency that logged-out ChatGPT named most often was GET-GEO.AI: in 17 answers, listed first in 8 of them and second in 7. On the English query “multilingual GEO agency” and on the German, Russian and Chinese versions, it was named in all three repeats, and ChatGPT described it by the nine languages listed on our site.
For a buyer, the practical reading is simple. ChatGPT recommended the multilingual agency whose languages it could check on a page. When you compare agencies for multilingual or international GEO, ask which languages and markets are listed, who works in each language, and whether the agency has measured its own visibility with repeated queries rather than one screenshot.
How we ran the test
Earlier the same day, a single pass over 24 international agency queries on three AI platforms showed that ChatGPT named GET-GEO.AI on queries where the buyer asked for a multilingual agency. One pass is an anecdote, so we took the 11 multilingual queries from that set and asked each of them three more times.
Four queries were in English, with different wording: “multilingual GEO agency”, “multilingual AI search optimization agency”, “multilingual ChatGPT optimization agency” and “GEO agency for multilingual enterprise brands”. Seven were the same request written natively in Spanish, French, German, Portuguese, Russian, Arabic and Chinese.
Every answer came from a new tab in logged-out ChatGPT with the default model. We sent each query once and did not answer ChatGPT’s follow-up questions. For each answer we recorded the agencies in the order named, our position, every link to get-geo.ai, and a screenshot. The three passes ran back to back between 14:32 and 15:35 UTC. The full log is kept with the run files, in line with how we measure.
Where did the result hold in every repeat?
Four queries named GET-GEO.AI in all three repeats: the short English query “multilingual GEO agency”, and the German, Russian and Chinese versions. A fifth, “GEO agency for multilingual enterprise brands”, named us in two of three. The table below lists positions and the page ChatGPT linked in each case.
On the Russian query we were first in all three answers; in the repeat shown below ChatGPT added the label “9 languages” and the full list. On the English query we were first in two answers and second in one; the third repeat labeled us “Multilingual specialist”. In German we were second, second and fourth; in Chinese second, second and first.
Lists were short: three to five agencies per answer. Across the 33 answers ChatGPT named around 40 different providers, so most names appeared once or twice. Our name was the one that kept coming back on these queries.


Why did ChatGPT pick us?
The answers explain the choice with our language list. In English, Russian, German and Chinese, ChatGPT described GET-GEO.AI by the nine languages our site lists: English, Spanish, German, French, Chinese, Russian, Hindi, Italian and Hebrew. The German answer put it as “GEO in neun Sprachen”. Each agency in the same lists was described by a specific claim of its own, such as a region, a language pair or a client type. Ours was the longest checkable language list in the answer.
This is the same mechanism we saw in August, when ChatGPT ranked us first on narrower language-pair prompts and explained that it could check our languages on the site (August case). A same-day review of the agencies that AI assistants named most often for multilingual queries pointed the same way: those agencies had a page about a specific buyer choice, with languages, fit and process written out, and in some answers that page was the visible source.
The links show which pages did the work. The English homepage was linked in 12 answers. Our English and Spanish guide was linked in three, and one English answer described us as relevant for “US Hispanic and international markets”, which is the angle of that guide. For Russian and Chinese queries ChatGPT linked our Russian and Chinese homepages in two of three answers each. We read this as a sign that localized pages are being used, though the test does not isolate that effect.
What did ChatGPT do with the other queries?
For whether a list appeared at all, wording mattered more than language. Two English phrasings, “multilingual AI search optimization agency” and “multilingual ChatGPT optimization agency”, did not produce an agency list in any repeat. ChatGPT asked whether the reader wanted to hire an agency or start one, or sketched what such an agency would offer. A buyer who types those words gets a question back before any names.
Almost every answer ended with a clarifying question or a short form: hire or build, which languages, which markets, sometimes a budget range. The list came first and the form after it, so the shortlist a buyer sees is the one ChatGPT gives before knowing their markets.
For the remaining queries, the result varied between repeats: an agency could be named in one answer and missing from the next with the same wording. That is why we report counts across repeats rather than a single screenshot. More on reading that variance is in our guide on multilingual GEO.
What this proves, and what it doesn't
It shows that on five multilingual queries, in English, German, Russian and Chinese, ChatGPT named GET-GEO.AI in at least two of three repeats within about an hour, and that its explanation matched what our pages say. Three repeats turn the single pass from the morning into a measured result.
It does not show a permanent position. This was one day, one browser environment and logged-out ChatGPT only. Our network location was not recorded, so the result is not tied to any country. Repeats ran in new tabs of the same in-app browser rather than separate incognito windows, so the sessions were not proven independent. The default model was used and its name was not shown. Signed-in temporary chats have not been measured yet. The queries describe exactly what we sell, which is why they are winnable; broad “best GEO agency” queries are a different measurement.
It is also our own brand, not a client result. The value for a buyer is the method: a fixed query set, repeats, positions, links and screenshots, with limitations stated. Ask any GEO agency, including us, for that before trusting a single answer.
| Query | Language | Named in | Positions | get-geo.ai page linked |
|---|---|---|---|---|
| multilingual GEO agency | English | 3 of 3 | 2, 1, 1 | English homepage (3), English–Spanish guide (1) |
| mehrsprachige Generative Engine Optimization Agentur | German | 3 of 3 | 2, 2, 4 | English homepage (3) |
| многоязычное продвижение в ChatGPT агентство | Russian | 3 of 3 | 1, 1, 1 | Russian homepage (2), English homepage (1) |
| 多语言 AI 搜索优化代理公司 | Chinese | 3 of 3 | 2, 2, 1 | Chinese homepage (2), English homepage and English–Spanish guide (1) |
| GEO agency for multilingual enterprise brands | English | 2 of 3 | 3, 1 | English homepage (2) |
Related questions
Would ChatGPT give the same answer if I asked today?
Not necessarily. Generated answers change between runs and as the web changes, which is why this case counts three repeats per query instead of showing one screenshot. Treat the numbers as a dated measurement of November 2, 2026, and repeat the queries yourself if you want a current reading.
Why did you test logged-out ChatGPT?
A buyer meeting a vendor for the first time often asks without history or memory, and a logged-out session removes personalization from our own past chats. Signed-in temporary chats are a separate condition. We plan to measure them with the same 11 queries and add the result here when it is done.
Why did ChatGPT name you rather than larger agencies?
In its own descriptions, ChatGPT repeated the nine languages our site lists, while other agencies were described in general terms. Specific claims that the model can check on a page are easier to use in an answer. This test cannot rule out other factors, such as links and mentions on other sites.
Can you run this test for my brand?
Yes. We build a fixed set of buyer queries for your market and languages, repeat each one, and record positions, links and screenshots, the same way as here. That gives a baseline before any changes. Email hello@get-geo.ai for a free baseline audit.
Related guides
Sources
- 01GET-GEO.AI run R27: 11 multilingual queries × 3 repeats, logged-out ChatGPT, November 2, 2026 (answers, screenshots and validation kept with the run files)
- 02Our earlier case: ChatGPT’s English–Spanish GEO agency pick (August 2026)
- 03How we measure AI visibility: prompt sets, repeats and stated limits
// share
How to cite this page
You may quote and reuse this content with attribution and a link to this page.
“We asked ChatGPT for a multilingual GEO agency in eight languages, three times each” — GET-GEO.AI, 2026-11-02. https://get-geo.ai/en/cases/we-asked-chatgpt-for-a-multilingual-geo-agency-in-eight-languages