Is My Website AI-Ready? 7 Signs ChatGPT Can't Find You
7 concrete signs ChatGPT, Perplexity, and AI Overviews can't find or cite your website — and the $19 Fix Report that scores the gaps with copy-ready edits.
A website is not AI-ready when ChatGPT cannot state what you sell, fetch the page, or verify a fact. Seven signs cover that gap: no plain offering, JavaScript-only content, no FAQ, blocked AI crawlers, missing JSON-LD, no llms.txt, and no credentials, numbered reviews, or featured-on listings — the Trust Signals slice of E-E-A-T. The $19 Fix Report scores those gaps and returns copy-ready edits.
7 Signs Your Website Isn't AI-Ready
- Your homepage never states what you sell in a plain sentence — visitors have to infer it from images or navigation labels.
- Key content loads only via JavaScript, so a non-browser fetch sees a blank or near-empty page.
- There's no FAQ section, and no page directly answers the questions people ask before buying.
- robots.txt blocks OAI-SearchBot or PerplexityBot. Blocking GPTBot is a training opt-out, not a ChatGPT miss; the token split is in Why ChatGPT and Perplexity Can't Find You.
- The site has no JSON-LD — no Organization, LocalBusiness, or FAQPage schema a machine could verify.
- There's no llms.txt file summarizing what the site is and where the useful pages live.
- Credentials, numbered reviews, or featured-on listings are mentioned nowhere a crawler can find them.
| Sign | Pillar it affects |
|---|---|
| No plain-language statement of what you sell | Business clarity |
| Content hidden behind JavaScript | Technical access |
| No FAQ or direct answers | Answer readiness |
| robots.txt blocks OAI-SearchBot or PerplexityBot | Technical access |
| No JSON-LD anywhere | Structured data |
| No llms.txt file | Technical access |
| No visible credentials, numbered reviews, or featured-on listings | Trust signals |
What each sign looks like on a real site
These are symptoms you can spot without a crawler log. The category they sit in is AI search readiness: can an assistant find the page, understand the offering, and verify a fact. How ChatGPT and Perplexity actually retrieve HTML (Bing, live fetchers, named robots.txt tokens) is a different post: Why ChatGPT and Perplexity Can't Find You.
The homepage never states what you sell
Open the homepage and read the first screen as if you had never heard of the company. If the offering is only in a logo, a mood-board hero, or a nav label like "Solutions," a model has nothing to quote. Business Clarity is the check for a plain sentence that names the category, the customer, and the outcome. A visitor can infer from photos. A fetch cannot.
Key content loads only through JavaScript
If the menu, services, or FAQ appear only after scripts run, a non-browser fetch gets a shell. Disable JavaScript and reload: whatever vanishes is what most AI crawlers never see. That is a Technical Access miss, not a design miss. Ranking in Google does not prove the same HTML is visible to ChatGPT-User or Perplexity-User.
There is no FAQ or direct-answer page
Assistants extract question-and-answer pairs more cleanly than dense brochure copy. If no page asks the questions buyers actually type, and no heading is followed by a self-contained answer, there is little to quote. Answer Readiness scores that pattern. FAQPage markup only helps when the questions are visible on the page.
robots.txt blocks or throttles AI crawlers
A leftover Disallow from the 2023 block-everything wave is the usual cause. Training tokens and retrieval tokens are not the same file rule. Blocking GPTBot is a training opt-out; it does not, by itself, explain a ChatGPT search miss. The token split is documented on GPTBot vs ChatGPT-User. Perplexity's index bot is PerplexityBot.
There is no JSON-LD a machine can verify
If view-source has no Organization, LocalBusiness, or FAQPage block, the page is asking the model to guess name, address, and offering from prose. Structured Data is the pillar that parses JSON-LD and checks it against visible copy. Markup that does not match the page is not a shortcut.
There is no llms.txt at the domain root
Missing llms.txt is a Technical Access finding: the file is a curated map, not robots.txt and not a sitemap. Consumer answer engines have not publicly confirmed they use it to pick citations. It is still a cheap, checkable gap, and the how-to is What is llms.txt.
Credentials, reviews, and featured-on listings are missing
If license numbers, numbered reviews, and featured-on listings exist only as images or widgets, a crawler cannot verify them. That is the Trust Signals slice of E-E-A-T the Fix Report scores: credentials, numbered reviews, featured-on listings. Authorship is not a scored Trust check. A tenure claim in a tagline, with nothing to verify, is not a substitute. See Trust Signals.
Why This Happens
Each sign traces back to one of six areas: business clarity, answer readiness, technical access, structured data, trust signals, and competitive visibility. A site can look finished in a browser and still fail a fetch. How Answer Patch weights those six areas is the methodology. The twelve-point DIY is the AI search readiness checklist. For how ChatGPT search picks pages, see how to optimize a site for ChatGPT search.