ObservatoryVisibilitas

Data

Updated Published

In short. Every report publishes its complete raw dataset: prompts, full answers, cited sources, counts. Licensed CC BY 4.0.

01Format

Each dataset includes: prompts.csv (the complete list of prompts used in the report), answers.jsonl (full text of every answer, by engine and date), mentions.csv (counts per entity, after disambiguation) and a README with the record layout and the methodology version used.

The prompt list of a report is published before its collection begins. Announced: prompts.csv for Who AI engines name in artificial intelligence — US, November 2026.

Datasets appear on this page as each report is published.

What every report dataset containsEvery dataset includes four files: prompts.csv with the complete list of the prompts used; answers.jsonl with full text, cited sources, engine and date; mentions.csv with counts per entity after disambiguation; a README with the record layout and the methodology version used. All of it is published under a CC BY 4.0 license.EVERY DATASET INCLUDESCSVprompts.csvcomplete list of the prompts used in the reportJSONLanswers.jsonlfull text · cited sources · engine · dateCSVmentions.csvcounts per entity, after disambiguationTEXTREADMErecord layout and methodology version usedCC BY 4.0 licensefree reuse, commercial too, with credit

Structure stated in the “Format” section above. The license is CC BY 4.0: datasets appear on this page as each report is published.