Readiness
The grade on your own site: whether engines can reach it, parse it, trust it, and whether they actually cite it. Four levels and four named tiers.
Readiness is the lens pointed at you rather than at the engines. It grades your own site on the four things that decide whether an engine can quote you, and it names the result so you can repeat it in a meeting.
What you'll learn
- The four levels, in the order they fail
- The four tiers, and why observed behaviour caps the grade
- What to fix first
The four levels
They fail bottom up. A blocked crawler makes perfect markup irrelevant.
| Level | The question it answers |
|---|---|
| 1. Access and crawlability | Can AI engines reach and read your site? |
| 2. Structure and machine readability | Can AI parse and extract your content? |
| 3. Authority and trust signals | Does AI have reason to trust you? |
| 4. Live AI citations | Is AI recommending you today? |
Access is about which crawlers you allow. The crawlers that fetch pages in order to cite them are not the same as the ones that gather training data, and they have different names. Blocking the first group costs you answers. Blocking the second is a licensing decision, not a fault, and the grade treats it as one.
Structure is about extraction, because engines take passages rather than pages. Titles, headings, canonical URLs, entity markup, and weighted highest of all: whether the content exists without JavaScript. No major AI crawler except Gemini and Applebot executes it, so a client-rendered page is often simply absent.
Authority is the tiebreaker between similar businesses. Links tying your entity to its other profiles, machine-readable contact details, review markup, freshness dates, and a page that says who is behind the business.
Citations is the only level that looks at real answers rather than at your pages.
The four tiers
| Tier | What it means |
|---|---|
| Invisible | AI answers do not surface you. Engines cannot find a clear, liftable answer on your site |
| Emerging | The foundations are forming, but engines still prefer other sources when they answer |
| Cited | Engines quote you on some prompts. The work now is widening which prompts you win |
| Preferred | Engines reach for you first on your core prompts. Defend it and extend the set |
The vocabulary is deliberately about being cited rather than about ranking.
Observed behaviour caps the grade
This is the rule that keeps readiness honest. Your page-side score can only claim so much once real engine answers are in evidence:
- Never cited across a real sample, and the grade is capped at Emerging however good the markup is.
- Cited only rarely, and the grade is capped at Cited rather than Preferred.
A minimum number of real checks is required before observation is allowed to cap anything, so one unlucky check cannot brand you invisible. When a cap applies, the panel says which observation applied it.
This is what separates a readiness grade from an on-page checklist. A checklist congratulates a tidy site the engines ignore. This one cannot.
What to fix first
Work bottom up, and stop at the first real failure. Most brands find their first significant win at level 1
or level 2, and it is usually one of three things: a crawler blocked in robots.txt, content that only exists
after JavaScript runs, or a page with no extractable answer on it at all.
Try this in Outercite
Ask the assistant "Is my website ready for AI answers?" The panel grades the four levels, names your tier, and tells you whether observed behaviour capped it.
Related
The six-level stack
These four levels plus the two above them: the change you shipped, and the proof it moved.
Changes
Where a fix to one of these levels goes to be tracked and proven.
Why is my citation rate low?
The diagnostic walkthrough.
Last updated on