AI Search Readiness guide · Chapter 4

Entity and page markup: saying clearly what the site is

Assistants and search engines identify a business from consistent names and valid structured data. Three places have to agree.

One name, everywhere

The title, the Open Graph title and the JSON-LD name should name the business the same way. "Acme", "Acme Widgets" and "acme.app" are three entities to a parser. The audit compares the leading segment of the title, the Open Graph title and the name declared on Organization, WebSite or SoftwareApplication markup, and reports a mismatch with the names it found.

The H1 is deliberately not part of the comparison. A headline is allowed to be a headline; the brand belongs in the title and the markup.

JSON-LD that parses and matches

Every JSON-LD block on a page is parsed. Invalid JSON is reported with the block. An Organization or SoftwareApplication with a name and url on the home page is the minimum; an Article type on blog posts helps the page type. FAQPage markup only counts when the questions appear as visible text on the page; the audit matches each question against the body and reports how many were found.

The repair can add Organization JSON-LD from the observed name and url only. It never invents addresses, founders or ratings, because markup that states things the page does not is a liability, not a signal.

Headings and the answer lead

One H1, sections under H2s, and a first paragraph that says what the site is in plain words. The audit reads the first 120 words of the main content and checks that they name the entity and form a sentence about it. The questions layer drafts that paragraph from your own description when it is missing.

Checks in this chapter

Structured data · Entity clarity · FAQ markup · Page type markup · Heading structure

Run these on my site