AI Readiness checks
Every check RankX AI's AI Readiness actually runs, with its category and its weight, generated from the product so it cannot list one that is not run.
Every AI Readiness check RankX AI runs, with the category it scores and the weight it carries when it fires. Generated from the product's own check registry, so this page cannot list a check the product does not run.
That is not a hypothetical safeguard. The plan that commissioned this page listed a check the product does not ship, and described one of the five categories as unscored when all five score. A list typed from a plan document is a list of what someone intended to build.
Every check
| Check | Category | Weight |
|---|---|---|
robots_blanket_disallow | Can AI reach you? | 30 |
ai_crawler_blocked | Can AI reach you? | 20 |
content_requires_js | Can AI reach you? | 18 |
commercial_noindex | Can AI reach you? | 12 |
snippet_controls_restrictive | Can AI reach you? | 12 |
llms_txt_uninformative | Can AI read you? | 10 |
llms_txt_absent | Can AI read you? | 8 |
facts_in_pdf | Can AI read you? | 5 |
entity_schema_absent | Does AI know what you are? | 16 |
service_schema_absent | Does AI know what you are? | 14 |
entity_sameas_missing | Does AI know what you are? | 9 |
pricing_page_missing | Can AI answer buyer questions? | 15 |
comparison_page_missing | Can AI answer buyer questions? | 12 |
pricing_not_parseable | Can AI answer buyer questions? | 11 |
integrations_page_missing | Can AI answer buyer questions? | 10 |
answer_first_absent | Can AI answer buyer questions? | 9 |
faq_schema_absent | Can AI answer buyer questions? | 4 |
no_review_footprint | Does AI trust you? | 8 |
author_signals_absent | Does AI trust you? | 7 |
stale_commercial_pages | Does AI trust you? | 7 |
page_dates_missing | Does AI trust you? | 6 |
page_dates_inconsistent | Does AI trust you? | 5 |
stale_dated_title | Does AI trust you? | 4 |
The table above is generated from RankX AI itself rather than written by hand, as at 19 August 2026 (source 377ef780). When the product changes, this page changes with it.
The five categories
They run cause to effect, which is also the order to fix them in.
| Category | The question it asks |
|---|---|
| Can AI reach you? | Are the assistants allowed to read your site at all? |
| Can AI read you? | When they do read it, can they tell what your business is? |
| Does AI know what you are? | Is what you sell written down in a form a machine can quote? |
| Can AI answer buyer questions? | When a buyer asks what you cost or how you compare, is there an answer to find? |
| Does AI trust you? | Is there anything beyond your own website vouching for you? |
All five are scored. A site the assistants cannot reach makes every other category a measurement of nothing, which is why the reach findings are worth fixing first regardless of what the weights say.
How the weight is used
A weight is what a check costs its category's score when it fires. Two rules change how that plays out, and both are on the AI Readiness page:
The overall score is your weakest category, not an average. So passing more easy checks cannot raise it. Only fixing the worst thing can.
A check that could not run moves the score in neither direction. It leaves both the numerator and the denominator, and is counted separately so the coverage gap stays visible.
A finding can also carry a weight lower than the table's default when the specific case is less serious than the check in general. The clearest example: a robots.txt that blocks only training crawlers costs far less than one that blocks a search crawler, because a deliberate training opt-out does not affect whether you appear in answers today.
Crawler checks are per purpose
The reach category checks fourteen crawler tokens across ChatGPT, Claude, Perplexity, Google, Apple, Meta, Amazon, TikTok and Common Crawl, and it distinguishes what each is for:
| Purpose | Blocking it |
|---|---|
| Search | Removes you from that assistant's answers. This is the one that costs you |
| Training | Keeps your content out of a future model. Does not affect answers today |
| User fetch | Stops the assistant fetching a page a person explicitly asked it to read |
Blocking training crawlers is a legitimate choice and RankX AI scores it as one. Blocking a search crawler is the finding that makes every other number on your account meaningless.
What the table does not carry
The verdict logic. Each check decides for itself what counts as passed, warned or failed, and several have thresholds that are not usefully summarised in a column.
The remedy. Findings carry their own remedy in the product, and it differs by what RankX AI can reach: some it can apply on a connected WordPress site, some it generates for you to place, and some it hands off.
Anything about your site. This is the catalogue. Your findings are on the AI Readiness screen.
Where to go next
- AI Readiness, for the scoring rules and the coverage floor.
- The Website Audit issue reference, which is the technical crawl rather than the machine-legibility scorecard.
- Null is not zero, for why an unrun check is not a pass.