The page's content only exists after JavaScript runs
Why it matters
AI crawlers do not execute JavaScript. GPTBot, ClaudeBot and PerplexityBot read the raw HTML and nothing else, so a page whose copy is assembled in the browser is a blank page to every answer engine except Google's.
What to do about it
Server-render the content, pre-render it at build time, or ship the copy in the initial HTML. Whatever the framework, the test is the same: view source and check the words are there.
This one is reported rather than applied. It needs a judgement a person has to make, and applying it automatically would be a change to your site that nobody decided on.
How it is weighted
These are the values the scoring actually uses, not a description of them. Priority is derived from all three rather than chosen, which is why a high-impact, low-effort finding surfaces above a severe one that would take a quarter to fix.
- Severity
- critical
- Impact
- 90% of the weight this category can carry
- Effort
- 70%, where 100% is a project rather than a change
- Confidence
- 75% that a hit is a real problem rather than a pattern we misread
- Area
- AI answers (AEO and GEO)
Confidence under 80% is deliberate. This check is narrowed so it fires less often, because a wrong high-severity finding teaches you to discount the severe ones, which are the only ones that matter.
Other checks in AI answers (AEO and GEO)
Check your own site for this
The whole catalogue over a real crawl, in about four minutes, in your browser. No account and no card.