AI finding
A page lacks what an AI assistant needs to read, trust and quote it
AI Warning Severity: Medium
Part of the audit Page readiness for AI answers. Finding id: ai.readinessGaps.
What it costs you
Assistants quote pages they can read without running scripts, whose facts they can check and whose author and date they can name. A page that fails here is passed over for one that does not.
What to do
Send the main text in the HTML, repair the structured data, name author and date on articles, link to your sources, and state each fact the same way everywhere.
What it checks
- Main content readable without scripts
- Structured data is valid
- Author and date shown
- Sources cited on the page
- Brand details consistent across the site
How to fix it
Render the main text on the server
- Turn on server-side rendering or prerendering for the template of [url], so the text is in the HTML before any script runs.
How to check: curl -s [url] contains the main text.
Correct the structured data error
- Correct the error shown ([detail]) where the markup is generated: the template, a plugin or a tag manager.
How to check: Run a structured data validator on [url], then run this rule again.
Name author and date on the article
- Show the author and the date on the page.
- Add author, datePublished and dateModified to the Article markup.
How to check: Run this rule again after the next crawl.
Link claims to their sources
- Link the claims in the text, figures first, to the source they come from.
How to check: Run this rule again after the next crawl.
State each fact the same way everywhere
- Pick the right value and correct every page that states another one.
How to check: Run this rule again after the next crawl.
Thresholds
| Name | Default | Meaning |
|---|---|---|
min_source_text_share | 0.5 | Share of the rendered text (0 to 1) that must already be in the HTML source. |
min_words_for_sources | 300 | Words of main content an article needs before missing outside sources are reported. |