AI Readiness Score · free · no email

How ready is your site to be cited by AI?

See your site the way an AI engine does. Type your address and we fetch your homepage, robots.txt, llms.txt and sitemap.xml — the same public files a crawler can request — then score the eight signals that decide whether an engine can read you, work out who you are, and find something worth quoting. Usually about ten seconds.

Get your AI Readiness Score

This reads your site. It never asks ChatGPT, Gemini or Perplexity anything — putting real buyer questions to three engines and reading what comes back takes a person, and that is step two.

Want to know whether the engines actually name you? A person answers that

Two different questions

One reads your site. One reads what the engines say about you.

No tool that only reads your own site can tell you whether ChatGPT names you. Not this one, and not any of the others. The score above answers the first question; only a person can answer the second, and that one is free too.

ComparisonAI Readiness ScoreAI Visibility Report
What it readsYour homepage, robots.txt, llms.txt and sitemap.xmlChatGPT, Gemini and Perplexity answers
Who runs itA script, in this pageA person
How longAbout ten secondsTwo working days
What it costsNothing, and no emailNothing, and a work email so the report has somewhere to go
What it answersCan an engine read and quote youDo the engines name you, and what do they say

What each of the eight signals is worth

Eight signals, 100 points. The checks are objective and you can verify every one by hand; the weighting is our judgement about what moves citation, not a published standard.

  1. Structured data 20 points

    Schema.org markup states in JSON what your pages say in prose. It is the difference between an engine inferring what you are and being told.

    Fix: Add Organization (or LocalBusiness), Service and FAQPage JSON-LD — and keep it identical to the visible page.

  2. AI crawler access 15 points

    If robots.txt disallows the AI crawlers at the site root, no amount of good content will get you cited. We check eleven known agents.

    Fix: Decide deliberately which AI crawlers to allow. Blocking them all forfeits citation.

  3. Server-rendered content 15 points

    Most AI crawlers read the raw HTML and move on. We count the words readable without JavaScript, and flag pages that look like an empty app shell.

    Fix: Server-render the copy that matters, so it exists in the raw HTML.

  4. Entity clarity 15 points

    A title, a meta description, exactly one H1, a canonical URL and an og:title let an engine say precisely who you are.

    Fix: Fill the gaps, and keep the same name and one-line description everywhere you appear.

  5. llms.txt 10 points

    A plain-text summary written for AI crawlers, at /llms.txt. A young convention, an afternoon's work, and it hands engines a canonical description of your business.

    Fix: Publish a /llms.txt summarising who you are, what you sell and your key facts.

  6. Answers to buyer questions 10 points

    FAQ blocks answering how much, how long and is it worth it are among the most-quoted content on any site. FAQPage markup scores full marks; question-style headings score partial.

    Fix: Add an FAQ answering the questions buyers actually ask, marked up as FAQPage.

  7. Third-party corroboration 8 points

    sameAs links to your profiles elsewhere let an engine verify your claims. Naming you is a risk for the engine; corroboration lowers it.

    Fix: List your verified profiles (directories, LinkedIn, partner listings) in sameAs.

  8. Sitemap 7 points

    A readable /sitemap.xml, referenced from robots.txt, helps crawlers find everything worth reading.

    Fix: Publish /sitemap.xml and reference it from robots.txt.

A worked example, imperfections included: this site scores 95 out of 100 as of 25 July 2026 — structured data 15/20, AI crawler access 15/15, server-rendered content 15/15, entity clarity 15/15, llms.txt 10/10, buyer answers 10/10, corroboration 8/8, sitemap 7/7. The missing five points are real: our homepage carries no top-level Service node and no BreadcrumbList.

Before you hand over an email

Who reads the answers.

Schnalz.net is a Hong Kong digital marketing agency and software studio. GEO is our flagship service, and we build the web and native apps that the fixes usually land in.

Google Partner since 2018 — agency ID 2839403546, five-star. Check it in Google's own directory rather than taking our word for it.

Clients include Bullish, block.one, Hong Kong FinTech Week, Cristel, Expansys, PlayAsia, Cognita and Shinsegae.

Verify the Partner listing

AI Visibility Report

What a person can tell you that the score can't.

We put the questions your customers actually ask — in English and in Chinese where your market needs both — to ChatGPT, Gemini and Perplexity, read the answers, and write up what they said.

  • Presence — do the engines mention your brand at all for buyer questions in your category?
  • Narrative — when you do come up, what do they say about you?
  • Competitors — who is recommended ahead of you, and what are they given credit for?
  • Priority — the three fixes we would make first, in the order the answers suggest.
Why this is free

GEO is our flagship service. A report that shows our work is worth more to us than a gated PDF, so the findings are yours either way — including if you fix everything yourself. If the gaps are worth closing and you want help, we quote a fixed scope and price upfront.

Request the report

Two required fields. Two working days.

A person runs the questions and writes the report.

No mailing list, no spam. One report, then it's your move.

We use your email to send this report and to reply about it. Nothing else. Delivered through Formcarry. Privacy policy

How this works

Four steps, and you can stop after any of them.

Nothing here commits you to anything. Two of the four are free, and one of them doesn't involve us at all.

01

Score your site

Seconds, at the top of this page. No email, no PDF, no follow-up. If you never come back, you still leave with the number and the fix list.

02

Fix what you can yourself

Every signal that misses comes with its fix. Several are an afternoon's work — an llms.txt, a sitemap referenced from robots.txt, an FAQ marked up properly. Server-rendering is the one that can be bigger.

03

Get the human report

We put buyer questions to ChatGPT, Gemini and Perplexity and send you what they said about you, within two working days. Free, no obligation.

04

Decide

If the gap is worth closing, we scope it and price it upfront, before you commit. If it isn't worth closing yet, we'll tell you that instead.

Stop wherever the answers stop being useful. The findings are yours either way.

Fair questions

The fine print, plainly.

About the AI Readiness Score

What is the difference between the score and the report?
The AI Readiness Score reads your site. It fetches your homepage, robots.txt, llms.txt and sitemap.xml and scores eight signals out of 100 — the things that decide whether an AI engine can read you, work out who you are and find something worth quoting. It runs in the page and needs no email. The AI Visibility Report reads the engines. A person puts real buyer questions to ChatGPT, Gemini and Perplexity, records whether you were named and who was named instead, and writes it up. That takes two working days, and it is also free.
Does the score predict whether ChatGPT will name me?
No. The score never asks ChatGPT anything — it only reads your site. A site can score 90 and still not come up in an answer, because what an engine says also depends on what exists about you elsewhere and on how the question was asked. Whether you are named today is exactly what the human report is for.
How long does the score take, and do I have to give you an email address?
Usually about ten seconds, longer if your site is slow to answer. No email: the score is ungated. Type the address, read the result, leave. The report needs an email because that is where a report gets sent. There is no mailing list either way.
Who decided the eight signals and their weights?
We did. The eight checks are objective and reproducible — you can verify every one of them by hand on your own site — but the weighting is our judgement about what moves citation, not a published standard. Here is the whole split: structured data 20, AI crawler access 15, server-rendered content 15, entity clarity 15, llms.txt 10, answers to buyer questions 10, third-party corroboration through sameAs 8, sitemap 7. Eighty and above is strong, 60 to 79 is solid, 40 to 59 is patchy, below 40 is weak. Nothing here is proprietary, so you can argue with the method instead of trusting it.
What does the scorer do to my site, and what happens to the address I type?
It requests four public things — your homepage, /robots.txt, /llms.txt and /sitemap.xml — the same ones any crawler can request. It logs in to nothing, submits nothing and changes nothing, and it identifies itself as SchnalzAIReadinessBot. We don't store the address, it isn't added to any list, and nobody here is notified unless you send the request form.
Can I score a competitor's site?
Yes. The tool only reads public files, so it works on any site you can visit. There is a ceiling of about a dozen scores an hour from one connection, to stop abuse. If you want the comparison done properly, name the competitor on the report form and we will include what the engines say about them alongside what they say about you.

About the AI Visibility Report

Why don't I just ask ChatGPT myself?
You can, and you should — it takes ten seconds and it is worth doing. The catch is that one person asking once is a single sample. Answers move with phrasing, with language, with the region you ask from and with the day. We put a fixed set of buyer questions to ChatGPT, Gemini and Perplexity, in English and in Chinese where your market needs both, and read them side by side. That is the part that does not fit in ten seconds. It is still one careful reading rather than a permanent ranking — repeating it monthly is how you see movement.
How long does the report take, and what language does it arrive in?
Within two working days, by email, written by a person. We write it in the language of the page you requested it from — English, traditional Chinese or simplified Chinese.
Is it really free — what's the catch?
Really free, and there is no catch: this is how we show our work instead of describing it. GEO is our flagship service, so a page that proves we know the subject is worth more to us than a gated download. The findings are yours either way, including if you fix everything yourself.
What happens after I get the report?
One email with the report. One follow-up offering a 20-minute call. If you don't reply, that is the end of it — you are not on a mailing list and there is nothing to unsubscribe from.

About working together

What does it cost if I want you to do the work?
GEO work is fixed scope, quoted upfront — you see the price before anything starts. Google Ads management is tiered to your monthly ad spend. It is all set out on the pricing page.
What is GEO?
Generative Engine Optimization — making your brand the one AI assistants cite when they answer buyers' questions. It builds on classic SEO but optimizes for being quoted in an answer, not ranked in a list of links. Read the full guide
Rather talk it through?

Twenty minutes, no deck.

Bring your score if you have one. We'll tell you honestly whether GEO is worth your money — and if it isn't yet, what is. GEO work is fixed scope, quoted upfront, so you see the number before you commit.

Book a 20-min call Get the free report first

See the work: Bullish · Hong Kong FinTech Week