# 2026 AI Content Authority Benchmark methodology

Snapshot date: 2026-08-29

Cohort: ranks 1-25 in the [Welcome.AI 100](https://www.welcome.ai/ai-100), scored June 2026. The source list selects the cohort; Heath Squier's audit does not reuse its scores.

The audit makes unauthenticated HTTP requests to each official site, robots.txt, llms.txt, declared or conventional sitemap locations, and one discoverable article candidate. It records 12 binary, publicly observable signals: robots availability, sitemap availability, llms.txt, Organization schema, feed discovery, Article schema, visible author, publication date, modified date, external citations, canonical URL, and social image.

Denominators follow observability. Site-wide signals use all 25 attempted domains. Homepage markup signals use only homepages that returned inspectable 2xx HTML. Article signals use only the article candidates the crawler discovered and fetched.

A failed or blocked fetch is recorded as unobserved, not proof that a feature is absent. Article discovery samples one recent candidate found in sitemaps or homepage links; it does not exhaustively crawl each domain, and a candidate may be a content hub rather than a conventional editorial article. The score measures observable content infrastructure, not content quality, traffic, rankings, or inclusion in AI answers.