Each signal has a one-sentence definition and a concrete reason it matters for being cited by an AI in 2026.
01
llms.txt
A plain-text file at the root of your domain telling AI crawlers which content to read.
why · Without it, ChatGPT and Perplexity often default to your homepage and miss your real content.
02
JSON-LD enriched
Structured data (Article, Person, Organization, Product) embedded in your HTML.
why · LLMs use it to understand who wrote what, when, and on which entity. Without it, the page is anonymous.
03
FAQ schema
Schema.org markup that flags questions and their answers as a Q&A block.
why · AI Overviews pull short, structured answers from these blocks before plain prose.
04
Heading hierarchy
A clean H1 then H2/H3 ladder, no skipped levels, no decorative H1s.
why · LLMs parse documents through their outline. A broken outline means a broken understanding.
05
AI bot access
GPTBot, ClaudeBot, PerplexityBot and CCBot allowed in robots.txt.
why · Many sites accidentally block these bots while allowing Googlebot. Citation becomes impossible.
06
Semantic tags
main, article, section and nav elements that mark where the content that matters is.
why · LLMs strip menus and footers. Semantic tags tell them exactly which block to read and quote.
07
Freshness signals
Dated content: time elements, article:published_time, visible modification dates.
why · AI engines favour sources they can date. An undated page reads as potentially stale.
08
Substantial text
Enough visible, self-standing prose on the page, beyond menus and captions.
why · An AI cannot quote a site that says nothing. Thin pages never make it into an answer.
Sk:vr measures 23 checks in total. These eight matter most for GEO and are the most often missing in 2026. The free report walks through each one and names exactly what your page is missing.