CiteLens by ShiftNode Digital

Measurement protocol · v1.0

How CiteLens measures AI citations

CiteLens runs one controlled set of unbranded buyer questions across major answer engines, records the source URLs they return, and reports whether the audited domain appears in that evidence. Google Search AI Overviews are measured as a separate surface, never inferred from Gemini. Last reviewed 28 July 2026.

What does an AI visibility audit measure?

An AI visibility audit measures whether an answer engine names or cites a website when it answers questions potential buyers could ask. CiteLens counts a citation when the audited domain appears in a returned source URL. A mention without a source link is recorded separately. The result is a sample, not a census, search-volume estimate or promise about the next answer.

The protocol

  1. Retrieve the submitted public page, metadata, headings, links, structured data and crawler rules.
  2. Generate roughly fifteen plausible, unbranded category and buyer-intent questions from that page.
  3. Send the same controlled set to the engines included in the plan and retain the question, provider status and returned sources.
  4. Separate returned source evidence from the later interpretation, scores and prioritized recommendations.

Paid audits cover Perplexity Sonar Pro, Google Gemini 2.5 Pro, OpenAI GPT-5.1, xAI Grok 4.5 and Anthropic Claude Sonnet 4.5 with web retrieval through OpenRouter. These are model/API probes, not tests of the providers' consumer interfaces. Guest and Free audits use Perplexity. Paid plans include a separate US-English desktop SerpAPI observation of the Google Search AI Overview panel. SerpAPI may return an identical cached result for up to one hour. If no AI Overview panel appears, the result is measured as no appearance. Provider errors or unusable responses are unmeasured.

Access is eligibility, not a guarantee

OAI-SearchBot supports ChatGPT search discovery; GPTBot is associated with possible model training. Googlebot controls Google Search eligibility, including AI features in Search; Google-Extended does not control Search inclusion or ranking. Perplexity documents Perplexity-User as a user-requested fetcher that generally ignores robots.txt, so a matching rule is an informational signal rather than proof that retrieval failed. Allowing a crawler only removes an access barrier. It does not make a page relevant, authoritative, indexed or likely to be cited.

Primary provider documentation

Limitations

  • Answer engines are probabilistic and can change sources between identical runs.
  • The generated questions are a controlled sample, not observed demand data.
  • Provider outages and missing source payloads reduce the usable denominator.
  • CiteLens scores are not endorsed by any search or AI provider.
  • No schema, crawler rule, llms.txt file or page edit can guarantee ranking, citation or inclusion in Google AI Overviews.