Original data

AI visibility statistics

Figures to 15 October 2026. Updated at the start of each month.

AI visibility statistics are measurements of how easily AI assistants such as ChatGPT, Claude and Perplexity can reach, read and trust websites. These show how often sites block AI crawlers, hide their text behind JavaScript, or leave out the signals AI assistants use to quote a page.

They come from websites scanned with CitableReport in the 12 months to October 2026, and no website is named.

Could we audit the website at all?

Before any check can run, a crawler has to get in and find something to read. This is how often that worked:

Audited in full 81%
Turned away at the firewall 0%
Empty without JavaScript 14%
Too few pages to read 6%

Access

Share of websites that…

Block at least one AI search crawler in robots.txt 0%

Check 1.2

Block AI training crawlers, but let AI search crawlers in 7%

Check 1.2

Stop indexing or quoting on some key pages 0%

Check 1.7

Have key pages that need JavaScript to show their text 0%

Check 1.8

Have no XML sitemap 3%

Check 1.4

Publish an llms.txt file 48%

Check 5.3

Structure

Share of websites that…

Pages that open with a direct answer, on a typical website 60%

Check 2.3

Use question headings or an FAQ 38%

Check 2.4

Have structured data on most pages 66%

Check 2.8

Facts

Share of websites that…

Link to their sources in articles (of websites with articles) 18%

Check 3.4

Show no sign of being kept up to date 14%

Check 3.7

Identity

Share of websites that…

Have no Organization structured data 14%

Check 4.1

Name an author on articles (of websites with articles) 36%

Check 4.2

Have no Wikidata entry 83%

Check 4.7

Show no reviews, ratings or press mentions 69%

Check 4.8

How do the scores spread?

The median website scores 75 out of 100. This is how scores spread, in ten-point steps:

And how they fall across the four bands:

Citable 14%

Nearly there 79%

Patchy 7%

Invisible 0%

Bands: Citable is 80 and above, Nearly there 60 to 79, Patchy 40 to 59, Invisible below 40.

Where do these figures come from?

They come from websites people chose to check with CitableReport. That is not a random sample of the web: it leans towards sites whose owners care about AI visibility, so the real picture may be worse.

One scan per website. Each website counts once, by its latest scan in the 12 months to October 2026. Repeat scans after fixes are left out.
Websites we could not audit count where the reason is about the site: a firewall refusing our crawler, pages empty without JavaScript, or too few pages to read. Typos, offline sites and our own errors do not.
The other figures are from fully audited websites, and leave out any check that could not be judged on a site. Each is based on at least 10 websites.
Each site is counted under one heading per check. For example, a site whose robots.txt blocks GPTBot but allows OAI-SearchBot counts under "block AI training crawlers but let the AI search crawlers in", not under "block at least one AI search crawler".
The median score is the middle one: half the websites score higher and half lower, so a few unusual sites do not drag it up or down.
The checks are the same fixed rules as in every report, set out on the audit page; scores use today's scoring rules.
Frozen monthly, so a figure you quote stays the same for the month.

Can I use these figures?

Yes. Please quote them with a link to this page, for example: "CitableReport, AI visibility statistics to October 2026".

Could we audit the website at all?

Access

Structure

Facts

Identity

How do the scores spread?

Where do these figures come from?

Can I use these figures?

See where your website stands