Open dataThe numbers,
The numbers,
in the open.
Methodology published, datasets on request, and every figure states the number it was taken from.
Buyers no longer search. They ask. And they walk into the decision with the name the assistant gave them.
10 → 1
results in a Google list. One name in an AI answer.
24%
of brands match between the API answer and the answer inside the app.
7 vs 16
sources cited for the same question, API versus app.
Published studies.
Real brands, real answers.
3
studies
2,776
AI answers read one by one
01A map of the companies of PortugalWho assistants name when asked about companies in the market · 616 answers · August 2026
02Politics as a control groupThe same method on a topic anyone can independently verify · 2,160 answers · September 2026
03Radar GEO PortugalA monthly report on the GEO and AEO consulting market · 3 editions · June to August 2026
A monthly public report on the Portuguese GEO and AEO consulting market, built from weekly competitor monitoring. Three editions published, one per month.
The methodology is not a promise. It is published.
01
The question is fixed before the data
What counts as a mention, and what counts as being chosen, is written down before a single answer is read. Deciding afterwards is how a study finds what its author was hoping for.
02
Names that are also common words are counted by reading
In Portuguese, NOS is also "in the", ERA is also a verb, and Logo is also the word for "soon". Those are counted by reading the sentence, not by matching a string, and the exceptions are listed.
03
Published numbers come from code, not from a conversation
Every figure comes from a script anyone can run again, with the denominators defined in one place and the result in version control, so a later run shows the difference as a diff.
04
The denominator is always stated
A percentage without the number it was taken from is not a finding. Where a question could not be measured, it is said, and it is kept out of the ratio rather than counted as a zero.
Raw CSV.
Redo the math.
Every row carries the question, the assistant, the date, and whether the answer came from memory or after searching, so a disagreement can be settled by looking.