Data
CiteProof data
What we see across the sites connected to the platform, aggregated and anonymous. Every figure carries its sample size and date; below the minimum sample we publish nothing.
Updated November 14, 2026. Window: last 30 days. Sites with recent data: 4.
AI crawlers
From the AI crawler visits the plugin recorded in the last 30 days.
Among sites with bot rules set, how many block in robots.txt at least one crawler the AI engines use to answer (OAI-SearchBot, PerplexityBot, Claude-SearchBot…). A site that blocks it cannot be cited by that engine.
How many sites decided which AI crawlers to allow or block. The rest let everyone read, training bots included.
Share of AI crawler visits that landed on archives, tags, internal searches and 404 pages: crawl that never becomes a citation.
Share of AI crawler visits that got an error (404, 410, 5xx).
People who clicked through from ChatGPT, Perplexity, Claude or Gemini, per 100 AI crawler visits. How much of the crawl turns into a click.
AI crawler visits by company.
- OpenAI53.8%
- Anthropic42%
- Other2.1%
- Perplexity1.2%
- You.com0.8%
- Meta0.1%
n = 2,482 visits
AI crawler visits by page type.
- other38.2%
- home21.8%
- page17.2%
- post14.8%
- singular3.3%
- tag1.7%
- category1.4%
- archive1.1%
n = 2,482 visits
Human visits from AI assistants by platform.
Content
From the snapshot the plugin sends with every sync: AEO scores, internal links, indexing.
Median of the average AEO score of sites with at least one scored page (0–100: short answer, FAQ, concrete facts, author, structure).
Share of scored pages with an AEO score under 55.
Sites with at least one published page no other page links to.
Sites that notify Bing and the other IndexNow engines on every publish or update.
Sites serving the XML sitemap generated by the plugin.
Brand and citations
From the fact-sheet consistency scan and completed audits.
Sites where an AI crawler, or a visitor sent by an assistant, requested at least one page that does not exist.
Among sites with a fact sheet, how many print at least one fact (phone, address, name, year) that differs from the declared one.
Median, across sites with a completed audit, of the share of questions in which the site was cited in its latest audit.
By sector
The same figures, sector by sector. A sector appears only once at least one figure passes the threshold (20 sites); the others stay below it and are not published.
No sector has reached the threshold yet: at least 20 connected sites with recent data per sector are needed.
How they are computed
- Source: the sites connected to the platform with the plugin active. Only sites with data from the last 7 days count; visits cover the last 30 days.
- Anonymity: only counts enter the computation. No domain, URL, title or customer data is aggregated or published.
- Thresholds: a percentage over sites needs at least 20 sites, one over visits at least 1,000 visits, one over pages at least 200 pages. Below the threshold the figure is not published.
- Cadence: recomputed once a day. The sample is CiteProof's customers, not a representative sample of the web: read these figures as orders of magnitude.
JSON version for automated citation: https://citeproof.co/api/public/benchmarks