Measures · Italy and Spain · October 18, 2026
Same question: in 12 cases out of 35 ChatGPT asks back, Perplexity gives names
- 12
- questions out of 35 in which ChatGPT asks back and names no one, and Perplexity gives names
- 6
- questions out of 23 in which both give names, but none in common
- 23
- names out of 121 given by ChatGPT that Perplexity also gives, to the same question
The first row is ChatGPT asking instead of naming. The other two say how alike the names are when both give some.
Source 40 questions from Mador's measurements, taken between 3 August and 15 October 2026 in 11 areas of Italy and Spain, read by hand on 17 and 18 October 2026. The businesses are not named: they are not customers and did not ask to be here.
Someone asks which dentist to choose in Santa Cruz de Tenerife. Perplexity answers with two clinics and a line on why. ChatGPT, given the same question in the same measurement, answers with questions of its own: what kind of dentist you need, whether price or experience matters more, which part of town you are in.
Mador read in full the two assistants' answers to the same question, across the questions people ask when looking for a business in Italy and Spain. In 12 cases out of 35, about one in three, it went the way it did with the dentist: ChatGPT asks back, Perplexity gives names.
The reverse doesn't happen
When ChatGPT gives no names, it sometimes offers advice on how to choose instead of questions. Still no name. In these questions the reverse never happened, meaning Perplexity with no names and ChatGPT with a list.
When both give names
Even when both give a list, the lists have little in common. In 6 cases out of 23 they share no business at all. In the rest they share one or two, never more, surrounded by names one gives and the other does not.
Just under one in five
Of the 121 names ChatGPT gave, Perplexity gave 23: just under one in five. The others are heard only by whoever asks ChatGPT.
More than half the time, not one name in common
Putting both cases together, ChatGPT giving no names and the two giving different ones, in 18 questions out of 35 whoever asks one and whoever asks the other do not hear a single name in common.
Watching one assistant shows half the picture
If you only check ChatGPT, you see one part. Not showing up there does not mean you show up nowhere. Showing up there does not mean people on Perplexity find you. For whoever is asking, these are two different lists for the same city, and neither knows about the other.
Where to start
Mador's free scan puts your area's questions to ChatGPT and Perplexity, and within a few minutes tells you which of the two you show up on and which you don't. Monitor puts them to 12 assistants every two weeks, for 59 € a month.
What this does not tell you
These are 40 questions read by hand: enough to see the shape, too few to turn into a percentage for all of Italy or all of Spain. When ChatGPT asks which part of town you are in and you answer, it does give names: what counts here is the first answer, the one read by whoever doesn't reply. And there are more than two assistants.
Method
- 40 questions drawn at random from 1,394, each with the two answers from the same measurement, read in full: only real businesses count, not portals, public bodies or neighbourhoods.
- Measurements from 3 August to 15 October 2026, read on 17 and 18 October 2026.
- ChatGPT and Perplexity, the same two assistants as the free scan.
Measured by Mador on October 18, 2026, with this method.
Want to know what the assistants say about your business? The analysis is free and takes a minute.
Run a free scan