Anonymised excerpt generated by ./audit excerpt from a real CLI-backend sample measurement (2026-10-09). Perplexity is not measured on the CLI backend; paid orders use 4 engines and 3 samples. Pricing

Sample B: Destination A, [state]

Anonymised tables from a real measurement. Answers come from API or CLI backends and are proxies, not what a given traveller sees. Google AI Overviews and AI Mode are not measured.

Discovery mention rate

DestinationChatGPTGeminiClaude
Destination A12 of 30 (min–max 40–40%, 2 samples)18 of 30 (min–max 60–60%, 2 samples)19 of 30 (min–max 60–67%, 2 samples)
Competitor 112 of 30 (min–max 33–47%, 2 samples)20 of 30 (min–max 67–67%, 2 samples)17 of 30 (min–max 53–60%, 2 samples)
Competitor 212 of 30 (min–max 33–47%, 2 samples)19 of 30 (min–max 60–67%, 2 samples)19 of 30 (min–max 60–67%, 2 samples)
Competitor 320 of 30 (min–max 67–67%, 2 samples)9 of 30 (min–max 27–33%, 2 samples)14 of 30 (min–max 47–47%, 2 samples)
Competitor 413 of 30 (min–max 33–53%, 2 samples)22 of 30 (min–max 67–80%, 2 samples)16 of 30 (min–max 47–60%, 2 samples)
Competitor 516 of 30 (min–max 47–60%, 2 samples)16 of 30 (min–max 53–53%, 2 samples)15 of 30 (min–max 47–53%, 2 samples)

Official-site citation rate

ChatGPTthe official site cited in 4 of 42 answers with citations (min–max 10–10%, 2 samples)
Geminithe official site cited in 4 of 6 answers with citations (min–max 50–100%, 2 samples)
Claudethe official site cited in 8 of 23 answers with citations (min–max 33–36%, 2 samples)

Cited sources by class

other68
state-tourism18
official16
competitor-dmo11
ota6
media5

AI crawler access

BotPurposeAt '/'
GPTBottrainingallowed
OAI-SearchBotsearchallowed
ChatGPT-Useruser-fetchallowed
ClaudeBottrainingallowed
Claude-SearchBotsearchallowed
Claude-Useruser-fetchallowed
PerplexityBotsearchallowed
Perplexity-Useruser-fetchallowed
Google-Extendedtrainingallowed
Applebot-Extendedtrainingallowed
CCBottrainingallowed

Ticket titles

  1. Cover the trip types where other destinations are named more
  2. Add TouristDestination structured data
  3. Re-run these 25 prompts quarterly (measurement and maintenance)
  4. Optional: publish an llms.txt
  5. Review the most-cited hosts each quarter (measurement and maintenance)