How it works
What Sited checks, and how the score is built
Every Sited report is generated fresh from your website URL: a live crawl of your pages, checked against five scored components, then written up as a single evidence-labelled report. No login, no connected accounts, nothing installed on your site.
The five scored components
AI platform readiness (25%)
Whether the crawlers behind Google AI Overviews, ChatGPT, Perplexity, Gemini and Bing Copilot can reach your site (robots.txt rules), and whether your content is readable without running JavaScript.
Content quality & E-E-A-T (25%)
Whether your pages give an AI system something substantial to quote and attribute: depth, clear headings, question-style sections, author bylines, publication dates, titles, descriptions and image alt text.
Technical foundation (20%)
HTTPS, security headers, response time, mobile viewport, sitemap, canonical tags and pages that fail to load: the basics that decide whether your pages get indexed properly in the first place.
Structured data & schema (15%)
The machine-readable markup that tells an AI system what a page is: an organisation, a website, a service, an article, a breadcrumb trail.
Brand authority & entity presence (15%)
Whether your site identifies who runs it (Organization schema, links to your official profiles, a consistent brand name). We do not check off-site presence such as Wikipedia or press coverage, so this component is capped at 80 until that is verified.
How the GEO Readiness Score works
Each component starts at 100 and loses points only for problems we observe on your site, such as a blocked crawler, a missing sitemap or a page with no structured data. The components are then combined using the weights above into a single GEO Readiness Score out of 100. The score is calculated by fixed rules, not judgement, so the same site always gets the same score, and your report lists every point deducted and why. Anything we cannot see from a crawl, like off-site reputation or real-world page speed, is marked as not checked rather than guessed.
Observed vs. experimental
We separate what we can directly observe on your site from newer, less-established practices where the evidence is still emerging. Your report's Agent Readiness section, for example, is explicitly labelled experimental: the signals are real, but the link to AI agent behaviour isn't yet backed by published platform guidance, so it never affects your score. Read more on our Agent Readiness page.
See it for yourself
We publish a full, unedited sample report generated for wikipedia.org, so you can see the real output format before you pay for your own.
