← Help center

AI Visibility explained

AI Visibility shows how often AI search engines mention or cite your site when people ask the questions you care about, and which competitors they pick instead.

Which engines we check

  • Google AI Overviews
  • ChatGPT Search: ChatGPT with web search turned on, the mode that cites sources.
  • Claude, Gemini and Perplexity

Checks run automatically on a schedule set by your plan: every 14 days on Standard, 7 days on Pro and 5 days on Enterprise. AI Visibility is not included on the Free plan.

Prompts

A prompt is a question we put to each engine. CrawlTide builds prompts from your keywords, Search Console queries and competitor gaps, and you can add your own. Prompt limits are 75 on Standard, 150 on Pro and 300 on Enterprise. Prompts that stop being relevant are archived automatically so old questions don't pile up in your results.

What each answer status means

Mentioned + Cited
The answer names your brand and links to your site as a source.
Mentioned
The answer names your brand but does not link to your site.
Source only
Your site is linked as a source, but the answer text does not name you.
Not found
We got an answer and you are not in it.
Not measured
We asked, but the engine did not return an answer (timeout or provider error). These are left out of the score instead of counting against you.

How the AI Visibility score works

The score runs from 0 to 100 and only uses answers we actually received.

  • For each engine we look at how often you are mentioned and how often you are cited.
  • If you've added competitors, we also look at how often they appear. Without competitors this part is simply left out, rather than scored for or against you.
  • Each engine gets its own score, and the overall score is the average across the engines that returned answers. An engine we checked more often doesn't count for more.
  • If no answer mentions or cites you, the score is 0. Tracking more prompts never raises it on its own.

Past scores in the trend chart are recalculated with the same rules, so a change in the chart is a real change in what the engines said.

Coverage and partial runs

Next to the score you'll see coverage, for example "Measured 132 of 136 answers · 4 of 5 engines". Coverage tells you how complete the latest check was. If an engine didn't respond, the run is marked partial: the answers we did get are scored, and the missing ones show as Not measured. A new engine shows "Not measured yet" until its first check.

Improving your score

  • The AI Readiness tab is a checklist of what search and AI engines document they need: search crawlers allowed, no noindex, snippets allowed, and pages that load. It shows the exact pages behind each failing check. Blocking training crawlers like GPTBot does not count against you, and structured data, business pages and llms.txt are shown for information only.
  • AI Citation Readiness lists the prompts where a competitor is cited and you aren't, with the page that should answer each one.
  • Log Files shows whether AI crawlers like GPTBot and PerplexityBot visit your pages at all.
Still stuck? Contact us and we'll help.