Skip to main content
Every figure in FullReach AI counts stored answers. The answer view opens one of them, so you can read what a platform actually said. It opens from every list of answers: Recent answers on the Overview, a prompt’s answers, a source, a search, a product and an Action’s evidence.

​
The header

  • The platform, and the model it reported, when it reported one.
  • The market the prompt ran in.
  • The time of the collection.

​
The question and the answer

The view shows the prompt. When the prompt’s persona has a sentence, Asked as shows the sentence that went in front of the question. See Personas. The answer follows word for word, as a logged-out visitor saw it in the platform’s own app, in a fresh session with no history. See How answers are collected. Highlights mark every brand that the analysis found. How answers are analysed says how.

​
Brands in this answer

Each brand the answer names, with its two labels: A brand that nobody monitors carries the mark detected. It counts in no figure.

​
Searches, products and sources

  • Searched the web for lists the searches the platform ran before it answered, in order. “No search ran” is a measurement. A platform that does not report its searches says so. See Searches.
  • Products shown lists the product cards, in order, each with its price and first store. See Products.
  • Sources cited lists the cited pages in the platform’s order. The number beside a page is its place in this list. The answer’s own inline markers are not collected.
Previous and Next step through the other answers of the list that opened the view.

​
A failed answer

A platform can block a request or return nothing usable. That answer is a technical failure, and it counts in no figure. When Google shows no AI Overview for a question, the answer is “No AI Overview”. That is a fact about Google, not a failure, and it counts against no brand. Both show in a prompt’s history.