How to Check Whether Your Site Is Ready for AI Search — Yourself
12% — that's exactly how many sites ranking in Google's top also appear in ChatGPT and Perplexity answers for the same queries, according to Ahrefs. You can hold rankings for years, collect organic traffic, watch your meta tags — and stay invisible to AI search. Not because the site is bad, but because no one prepared it for GEO.
This article is a self-check checklist: go through it item by item and see for yourself what on your site is already ready for AI and what's missing. No vague advice like "just make good content" — only what actually affects AI visibility, each item with a time estimate. If you want the "why and how it works" explainer instead, that's in a separate guide, how to get into ChatGPT answers; here we check the site.
The GEO Check-up Map: Where to Start and in What Order
Before adding schema.org to every page at once, agree on priorities. GEO optimization is a layered process, and the order of the layers matters. Here's the breakdown by impact:
| Priority | What to do | Why it comes first | Time |
|---|---|---|---|
| 🔴 Critical | Open access to AI crawlers | If the bots can't get in, the rest is pointless | 5 min |
| 🔴 Critical | Schema.org: Organization + WebPage | Without it the AI doesn't know who you are or what the page is about | 15–30 min |
| 🟡 Important | FAQPage and Article markup | The format AI engines cite most willingly | 30–60 min |
| 🟡 Important | Content built for fact extraction | AI doesn't read — it pulls facts. No facts → nothing to cite | 2–4 h |
| 🟢 Later | E-E-A-T signals | About page, authorship, fresh dates | 1–2 h |
| 🟢 Later | External brand footprint | Directories, review sites, maps | A week+ |
| ⚪ Background | llms.txt, sitemap, speed | Technical tuning to amplify, not to start | As needed |
Got an hour — do the red. Got a day — add the yellow. Spread the rest across the week. Now each layer in detail.
Technical Factors: Open Access to AI Crawlers
This happens more often than you'd think: a site blocks ChatGPT's bots in robots.txt. Not on purpose — someone once grabbed a "standard" file off the internet that already disallowed GPTBot.
Every AI engine has its own crawler, and access for Google doesn't mean access for Perplexity.
Who Is Trying to Reach Your Site
| Crawler | Owner | Purpose |
|---|---|---|
| GPTBot | OpenAI | Collecting training data |
| OAI-SearchBot | OpenAI | Search with attribution and links — key for citation |
| Google-Extended | Access for AI Overviews | |
| PerplexityBot | Perplexity | Indexing for Perplexity answers |
| Claude-Web | Anthropic | Access for Claude |
| Applebot-Extended | Apple | Access for Apple Intelligence |
Official bot lists and rules: OpenAI, Perplexity, Google.
How to Check and What to Add
Open yoursite.com/robots.txt and look for Disallow rules for these bots. If User-agent: GPTBot has Disallow: / — access to AI search is closed. The minimum robots.txt snippet for GEO:
User-agent: GPTBot
Allow: /
User-agent: OAI-SearchBot
Allow: /Five minutes — and the first barrier is gone.
Schema.org: What to Mark Up First
If robots.txt is the entrance, Schema.org is the language your site uses to talk to the AI. Without markup, AI sees plain text; with markup, it distinguishes a company, an article, a guide, a product. Of the 800+ Schema.org types, only a handful matter for AI visibility.
The minimum — nothing works without it
- Organization — company name, URL, email, phone, social links, logo. The "real company" signal.
- WebPage — page type, language, publisher. Base context for the AI.
The boost — for pages that should appear in AI answers
| Schema | For what | What it gives the AI |
|---|---|---|
| Article | Articles and posts | Headline, date, author — "made then, by that expert" |
| FAQPage | Question-answer blocks | The most cited format: ChatGPT and AI search pull ready pairs directly |
| HowTo | Step-by-step guides | Citation of guides and "how to" answers |
| Product / Service | Stores and services | Price, availability, service area |
The format is JSON-LD, not Microdata or RDFa: JSON-LD is the de facto standard for AI engines. After adding it, always run the markup through the Schema.org validator — if it shows errors, the AI sees them too and ignores the schema. Time: 30 minutes for the base set, an hour for FAQ and HowTo.
Content the AI Can Actually Cite
The key difference: the AI is not a reader. It isn't there to judge your prose and the flow between paragraphs. Its job is to find concrete statements and decide whether they can be cited.
| Text variant | What the AI takes |
|---|---|
| "A wide range of quality digital-marketing services, helping businesses achieve impressive results with cutting-edge technology." | Nothing: no number, no fact, no date. |
| "The GEO / AEO audit in SEOFetcher checks 30+ parameters in 3–5 minutes, shows missing markup and gives recommendations — free, no sign-up." | Three facts: 30+ parameters, 3–5 minutes, free — each citable on its own. |
The GEO content rule
Every paragraph contains at least one machine-readable specific — a number, date, name, price, address, result. Paragraphs you could delete without losing meaning are ignored by AI.
Heading structure
Write H2–H3 headings as questions or concrete topics: not "Our advantages", but "How to check whether your site is indexed". AI looks for question → answer pairs: if the heading is a question and the block below is a direct answer, the page gets cited.
Formats AI handles most readily
- Numbered and bulleted lists
- Tables
- A takeaway after each section ("In short: …")
- FAQ blocks with real customer questions
Time: 2–4 hours on key pages. The most labor-intensive part of GEO optimization — and at the same time the most impactful.
E-E-A-T: What AI Checks Beyond Your Site
E-E-A-T — Experience, Expertise, Authoritativeness, Trustworthiness. Originally a Google framework for human quality raters. AI picked it up because choosing the wrong source is a reputational risk for it. What AI checks specifically:
- Does an About page exist and contain something beyond "a dynamic, growing team"? Are there real company details, an address, names of real people?
- Are article authors listed with full names and links to profiles — or is it "Admin" everywhere?
- Are publication and update dates present? AI favors pages updated within the last 3 months.
- External mentions — a verified business profile, map and directory listings, reviews on independent platforms, coverage in industry media. If nothing outside your own site mentions the company, authority is zero.
Buying links doesn't work for GEO — that's a classic SEO tool. What matters here is real presence in the information field: one mention from an industry expert on a respected publication outweighs a hundred marketplace links.
Tools to Check GEO Readiness Yourself
Everything above can be checked manually, but it takes an hour or two per site — and proportionally more for several.
Automated check
The SEOFetcher GEO / AEO audit — enter a URL and in 30 seconds get a report: JSON-LD markup, E-E-A-T signal level, AEO readiness. Free, no sign-up.
Check your site's AI readiness
The SEOFetcher GEO / AEO audit — JSON-LD, E-E-A-T, FAQPage, HowTo in 30 seconds. No sign-up.
Manual check, point by point
| What to check | Where |
|---|---|
| robots.txt | yoursite.com/robots.txt |
| Markup validity | validator.schema.org |
| AI crawler access | Google Search Console → Crawl stats |
| Indexing in Google | a site:yoursite.com query — if suspiciously few pages, look for a technical cause |
| llms.txt | yoursite.com/llms.txt — if 404, worth creating, takes 10 minutes |
The llms.txt file is the robots.txt equivalent for language models: a map of your key pages with short descriptions.
Monitoring the result
Once a month, ask ChatGPT and Perplexity 10–15 questions in your niche and track when your site appears and disappears. The first movement is usually visible 2–4 weeks after implementing GEO optimization.
Frequently asked questions
Where do I start if resources are minimal?
robots.txt (5 min) → confirm JSON-LD Organization is on the homepage (10 min) → add an FAQ block to your most important page (30 min) → rename 2–3 headings into questions (10 min). One hour and the base is ready.
How long until GEO optimization shows results?
2–4 weeks to the first citations, 2–3 months to a stable presence in AI results. Slower than classic SEO, but competition is still lower.
Does GEO replace SEO?
No. If a site is technically broken and not indexed, the AI won't find it. GEO and SEO work together: SEO provides the base, GEO brings AI traffic.
Which markup matters most for getting into AI answers?
The base is Organization and WebPage — without them the AI doesn't know who you are or what the page is about. Next is FAQPage, the most cited format, because ChatGPT and Perplexity pull ready question-answer pairs directly. All in JSON-LD, validated with the Schema.org validator.
Do I need an llms.txt file?
It's recommended. llms.txt is a map of key pages for language models, the robots.txt equivalent. It isn't mandatory, but it takes 10 minutes to create and helps the AI grasp your site structure faster. If yoursite.com/llms.txt returns 404, it's worth adding.
Check Your Site Right Now
Free GEO / AEO audit: JSON-LD, E-E-A-T, schema.org — no sign-up, takes 30 seconds.