Standards

An evidence block is a fixed seven-item module that presents evidence inside a page, not a footer disclaimer.

Updated 2026-11-09

canonical is a link tag declaring a page's preferred address, used to consolidate duplicate content signals.

Updated 2026-11-09

JSON-LD is a structured data format embedded in a page as JSON; it lowers comprehension cost, not a citation promise.

Updated 2026-11-09

llms.txt is a community proposal placing a Markdown content index for large models at the site root.

Updated 2026-11-09

ECI is the share of sampled claims that pass all six hops, showing whether evidence chains actually work.

Updated 2026-11-09

Evidence levels mark claim credibility from L1 to L5; cross-citing owned domains counts at most as L4.

Updated 2026-11-09

An entity is a person, organization, or concept identified by a fixed @id so AI can merge the subject across pages.

Updated 2026-11-09

Extractable means the first screen holds a conclusion sentence under 60 characters that still stands out of context.

Updated 2026-11-09

The three-click rule requires that any landing page reach a primary source in at most three clicks.

Updated 2026-11-09

The six hops are discoverable, openable, extractable, verifiable, attributable, and corroborable stages of verification.

Updated 2026-11-09

An evidence chain is a clickable verification path from a claim to a primary source and then to external corroboration.

Updated 2026-11-09

A primary answer means the core conclusion of an AI answer comes from this site, not merely as a supplementary source.

Updated 2026-11-09

An AI mention is a brand name appearing in an answer without a link to this site; it is not an endorsement.

Updated 2026-11-09

An AI citation is a clickable URL pointing to this site inside an AI answer; a name alone does not count as one.

Updated 2026-11-09

RAG is a model-side mechanism in which a large model retrieves external documents before generating an answer.

Updated 2026-11-09

SEO is the technical and editorial work of improving a page's visibility and position in traditional search results.

Updated 2026-11-09

LLMO shapes, through corpus and entity signals, how a large model represents a brand internally and what it generates.

Updated 2026-11-09

GAO is the goal of giving AI an accurate, positive, repeatable perception of a brand in answers and recommendations.

Updated 2026-11-09

AEO makes a page the preferred source for direct answer formats such as Q&A and AI summaries, using verifiable answers.

Updated 2026-11-09

GEO gets generative AI to cite and accurately describe your content. It targets citations, not search ranking.

Updated 2026-11-09

Evidence freshness is the share of evidence verified within its review cycle, with cycles set by evidence type.

Updated 2026-11-09

Broken evidence rate is the share of evidence links that fail, checked monthly in batches and reviewed by hand.

Updated 2026-11-09

Primary source coverage is the share of source entries pointing to original documents or datasets, counted monthly.

Updated 2026-11-09

ECI is the share of sampled claims that pass all six hops, sampled quarterly with the rules published.

Updated 2026-11-09

Wording accuracy rate is the share of mentions whose name, title, and organization match the wording baseline.

Updated 2026-11-09

Primary answer rate is the share of questions whose core conclusion comes from this site; judgment is manual and traced.

Updated 2026-11-09

Mention rate is the share of questions whose answers name the brand or person, with or without a link; not endorsement.

Updated 2026-11-09

Citation rate is the share of questions whose answers link clickably to this site; cited questions over total questions.

Updated 2026-11-09