When a traveller asks an AI trip planner where to go, is your destination named?
We ask ChatGPT, Claude, Gemini and Perplexity the questions travellers from your origin market ask ("weekend beach trip from Grand Rapids", "fall colour drives in the Midwest"), three times each. The report shows how often each engine names your destination, whether it cites your official site, which destinations, OTAs and blogs it names and cites instead, and whether AI search crawlers can read your site. You get 5 developer tickets and a one-page board summary.
Built for small and mid-size US destination marketing organisations: CVBs, county and city tourism departments and tourism councils.
Three anonymised sample excerpts
Tables from three real measurements of public US destination websites, run on 2026-10-09 with the CLI backend (ChatGPT, Claude and Gemini; 25 prompts × 2 samples per engine). Destination, competitor and site names are replaced. The full reports are not published.
Sample A
Mention rates against 5 competitor destinations, official-site citation rates, cited sources by class, AI crawler table, ticket titles.
What the report covers
- Mention rate vs competitor destinations: per engine, mean and min–max over samples, with the samples shown. A destination is named or not named; the exact matching sentence is kept for every count.
- Official-site citation rate and the top cited hosts by class (official, state tourism, competitor DMO, OTA, media, other).
- Prompt-level grid: each traveller question × engine, named yes/no per sample, official site cited yes/no, linked to the stored answer spans.
- AI crawler access: robots.txt rules per bot, separating training bots (GPTBot, ClaudeBot, Google-Extended, CCBot …) from search and user-fetch bots (OAI-SearchBot, Claude-SearchBot, PerplexityBot …), plus Content-Signal and meta robots.
- Structured data and sitemap: JSON-LD types found and sitemap counts.
- Plan-your-visit pages: raw-HTML word counts and presence checks.
- 5 developer tickets (Markdown and CSV) and a one-page board summary.
- Method and limits, including every item we could not measure and why.
Pricing
| Package | What you get | Price (USD) | |
|---|---|---|---|
| Baseline | 25 traveller prompts × 3 samples × 4 engines (300 answers), 1 origin market, up to 25 pages of your site read, up to 5 competitor destinations. Report (HTML + PDF), 5 developer tickets, 1-page board summary. Delivered 3 business days after we receive your intake answers. | 495 | Buy Baseline |
| Board Pack | Everything in Baseline with 2 origin markets (50 prompts × 3 samples × 4 engines, 600 answers), up to 40 pages read, a per-origin split, a competitor-destination benchmark table and a day-90 re-measure with the identical prompts and a delta table. Delivered 3 business days after intake; re-measure 90 days after delivery. | 1,250 | Buy Board Pack |
Pay by card or P-card through Stripe. At checkout, enter your destination and website. After payment we email the six intake questions from support@plainworktools.com. If your organisation needs a W-8BEN or a vendor form first, email us before buying.
Refund: full refund if the audit cannot run (your robots.txt disallows our crawler at
/, the site does not answer, or your site terms bar automated access). One free revision after
delivery. See the audit terms.
Limits we print in every report
- Answers come from API or CLI backends. They are proxies, not what chatgpt.com, claude.ai, gemini.google.com or perplexity.ai show a given traveller.
- Google AI Overviews and AI Mode are not measured. We make no automated Google requests.
- Gemini runs without web search, so it returns no citations.
- Answers vary between runs, so every rate shows its samples and its min–max.
- A blocked training bot, a missing schema type or a low word count is a readiness finding. It is not proof that a destination is invisible, and the report never says it caused a rate.
- No outcome is promised: the report measures where you stand today and what to fix.
What we never ask for
No logins, analytics, Search Console, CMS, Simpleview or Google Business Profile access. We read only your public
pages, obey your robots.txt and Crawl-delay, and identify our crawler as GeoAuditBot. We never
pretend to be an AI bot.