How to read the Answers screen

In a brand's workspace, the "Questions and answers" ("Sorular ve yanıtlar") section asks one question: what exactly did the AI say? The evidence for every number in the report is on the Answers (Yanıtlar) screen.

What is in a record?

Each record is one question asked to one engine one time. The record contains the question, the engine that answered, the answer text the engine gave and an identity label. The answer text is not edited; it is the answer as it was at the moment it was asked.

The identity label tells you how this answer feeds into the numbers. The two you will see most often are "Brand appeared in the answer" ("Marka yanıtta geçti") and "Measured · brand not mentioned" ("Ölçüldü · marka anılmadı"); both are measured results. Records labelled "Could not be measured · no answer received" ("Ölçülemedi · yanıt alınamadı"), on the other hand, are gaps on our side and count toward no rate. The full list of labels is in The difference between not measured and not mentioned.

If the answer cited sources, the number of sources captured is also shown in the record. If an answer's source information was not stored, this is stated separately; it does not mean "cited no sources".

Filters

You can narrow the list by these headings: Engine, Question type, Identity status, Audience context, Intent family and text search. Question type takes three values:

• Question asked without your name (Adınız geçmeden sorulan soru): the core questions of the visibility measurement.
• Control question, brand name in the question (Kontrol sorusu, marka adı soruda geçiyor): shows whether the AI recognises you.
• Question not counted (Sayıma katılmayan soru): other questions that are not part of the measurement.

The reason for this split is in Questions that include your name and questions that do not.

The list is not a measurement

When you apply a filter, the number of remaining records is not a rate. Official rates are calculated over the measurement's own groups and are shown in a separate box on the screen, read exactly as they appear in the published report; filters do not change them. For a selection that has no stored measurement, such as a text search or the intersection of two filters, the screen shows no rate and explains why.

Personas

The Personas (Personalar) screen in the same section shows the buyer profiles used when writing the questions, and looks at this question: who is asking, and what does the AI show them?