Pages Optimization
The issue queue for your own pages: severity, category, what the engine found, and the specific fix.
Pages Optimization is the defect tracker for AI citability. Pages scores your URLs; this screen lists every issue those audits found, ranked so a developer or writer can work top-down without triaging first.
Preview. The page audit crawler is not yet running, so this screen renders a sample issue queue rather than defects found on your site. See Product status.
Anatomy of an issue
Severity
| Level | Meaning |
|---|---|
| Critical | Actively costing you citations. Fix this week. |
| Warning | Measurably weakening the page. Fix this month. |
| Info | Worth doing when the page is next touched. |
Category
| Category | What it covers |
|---|---|
| AI Citability | Whether a passage can be lifted: structured data, direct answers, extractable formatting. |
| Authority & Trust | Whether the page deserves to be believed: references, freshness, depth, provenance. |
| AI Readability | Whether the page can be parsed at all: markup, hierarchy, content not hidden behind scripts. |
Page — the exact URL affected.
First detected — how long it has been open. An old Critical is a process problem, not a content problem.
What AI found — the observed behavior that produced the issue. For example: "The page answers six recurring pricing questions in prose, but no FAQPage structured data exists. Passage extractors fall back to competitor pages that mark identical questions up."
Suggested fix — the concrete change, specific enough to implement without interpretation: emit the markup, mirror the questions as headings, keep answers under sixty words.
Working the queue
Filter by severity, by category or by page, and search across issues.
Three ways to sequence the work, all defensible:
- By severity — the default. Critical first.
- By page — everything wrong with one URL, fixed in a single pull request. Efficient when you are already editing that template.
- By category — all structured-data issues at once. Efficient when the same fix repeats across pages, which it usually does.
The second and third are underrated. Ten Warnings sharing one root cause are one task, not ten.
After shipping
- Deploy the change.
- Re-analyze the page from Pages.
- Confirm the issue clears and the page's citability score moves.
- Wait for the engines to re-crawl before judging the visibility effect — days, not hours.
What tends to matter most
Patterns that recur across audits, roughly in order of payoff:
- Missing structured data on pages that already answer the questions in prose. The cheapest fix in the product, and consistently the most effective.
- Answers buried below the fold. Engines extract passages; a definition on line four beats the same definition after three hundred words of preamble.
- Comparison pages that avoid naming competitors. The engine is answering "X vs Y". A page that will not say Y is not a candidate.
- Stale dates and stale facts. Freshness is a trust signal and an easy one to lose.
- Content rendered only by client-side script. If a crawler cannot see it, it does not exist.
None of this is new to anyone who has done technical SEO. What is new is the feedback loop: the issue was raised because an engine visibly behaved that way on your prompts, and you can watch it change.