Testing what AI says about your park.
There is no dashboard for this yet. There is a set of questions, four assistants, and a spreadsheet. Ask the same things every month and the trend is plain, and every wrong answer points at a page to fix.
IN ONE PARAGRAPH
GEO is measured by asking. Pick ten prompts: five where a traveler would want your park recommended by town and features, and five about your park by name. Run them in ChatGPT, Gemini, Copilot, and Perplexity monthly, in a fresh session, and record whether you were named, what was said, what was wrong, and which page was cited. Named-in-recommendation count and factual-accuracy count are the two numbers. Every wrong fact maps to a source: the site, the profile, a directory, or a stale review. Every absence maps to an index gap or a missing entity. Fix, wait a month, ask again.
The prompt set
Recommendation prompts, in a traveler's words:
- RV park near [town] with 50-amp pull-through sites for a 45-foot fifth wheel.
- Pet-friendly campground near [attraction] with full hookups.
- Where to stay in an RV near [town] in [your best month].
- Monthly RV site near [town] with good WiFi for remote work.
- Campground with cabins near [attraction] for a family of four.
Name prompts:
- Tell me about [park].
- What is the longest site at [park] and what amperage does it have?
- Is [park] open in [winter month]? What are the hours?
- Does [park] allow dogs, and is there a fee?
- How far is [park] from [attraction]?
What to record
| Field | Example |
|---|---|
| Date, assistant, prompt | Sep 2026, ChatGPT, prompt 1 |
| Named? | Yes, second of three |
| What it said | "70-foot pull-throughs with 50 amp, open Apr to Oct" |
| Wrong facts | Season: we are year-round now |
| Source cited | Campendium listing |
| Fix | Update Campendium season |
What one looks like
A park's first three months, as recorded:
| Month | Named in recommendations (of 20) | Name prompts correct (of 20) | What changed |
|---|---|---|---|
| Launch | 3 | 9 | Site live, Bing verified, facts page up |
| Month 1 | 7 | 14 | Bing indexed; two directories corrected |
| Month 2 | 11 | 18 | Profile season fixed; reviews program started |
| Month 3 | 14 | 20 | Big-rig page added; Perplexity now cites it |
Twenty recommendation slots are five prompts times four assistants. The park went from named three times to fourteen, and every step came from a specific fix the answers pointed at.
A 20-site park moves faster, because there is less to fix and less competition for its specific facts. Cedar Bend RV Park went from zero to nine of twenty in two months on a Bing verification, a facts page, and one directory correction.
Turning answers into fixes
- Not named by ChatGPT or Copilot but named by Gemini: Bing index gap. Verify with Bing.
- Named but with a wrong fact: find the source it cited and correct it, then check the site says the same.
- Named vaguely, "offers full hookups in a scenic setting": the pages have no quotable sentences. Write the facts.
- "Listed on [platform]": the entity has no home. The platform page problem.
- "Check whether it is open": stale reviews or a season mismatch.
How to check your own
- Run the ten prompts in one assistant this week. Thirty minutes.
- Record named, said, wrong, cited.
- Fix the one thing the answers most agree on.
What we do about it
The Marketing band runs this set monthly across the four assistants and reports the two numbers with the fixes made. It is the report that shows GEO working or not, in the park's own town, in the words a traveler would use. What not to do, last.
Questions about this.
How do I know if AI is recommending my RV park?
Ask it. Run the same set of traveler prompts in ChatGPT, Gemini, Copilot, and Perplexity monthly and record whether you are named and what is said.
Is there a tool that tracks AI mentions for a campground?
Several are emerging, but for one park the manual monthly prompt set is more accurate and free. Ten prompts, four assistants, one spreadsheet.
Why do AI answers about my park change from day to day?
Assistants search at the moment of the question and phrase answers differently each time. Run prompts in fresh sessions and look at the monthly trend, not any single answer.
What is a good result for GEO?
Named in most recommendation prompts for your town and features, and every name prompt answered with correct facts. Most parks start at nearly zero, so any steady climb is the signal.
Want to know what AI says about your park?
Ask it, then call us. We run the prompt set, show you the before, and build the fixes. SEO, AEO, and GEO with us start at $250 a month.
Part of the RV park GEO guide. Previous: The booking-platform page problem. Next: What not to do. See also the SEO guide and the AEO guide.