Gate one of four
If a model can't open your site, nothing else you do matters.
The first gate is mechanical, and it is the one most often failed without anyone noticing. A setting written years ago, a security rule nobody reviewed, or a page that only exists once JavaScript has run: the result is the same. To a model, your site is blank.
- Pages a person can read
- 26 / 26
- Pages a model can read
- 4 / 26
- Words in the answer it needs
- 0 / 1,240
Why this one decides the rest
A person opening your site in a browser sees the finished page. A model does not. It fetches the raw response, follows whatever your server tells it, and reads what comes back. The two are often very different documents.
Nobody is notified when this breaks. There is no error, no warning in your inbox, and your traffic from people does not move. The only symptom is an absence: you stop appearing in answers you used to appear in.
- No access, no chance If the model can't get in, none of your content can be considered.
- One setting, big impact A single rule can hide your entire website from every model.
- Silent by default Many blocks don't create errors. They just return an empty page.
- Fixable in hours Most issues are simple to find and even simpler to fix.
What this looks like in your report
The panels below are a worked example, run against a roofing company in Columbus, Ohio we made up for the purpose. Yours arrives with your own answers, word for word.
Same page, two readers
The same address, fetched twice. A browser runs your page before showing it. A model reads what the server sent.
-
What a visitor sees
1,240
-
What a model receives
0
| What a visitor sees | 1,240 words |
|---|---|
| What a model receives | 0 words |
Gate one
-
The rule that tells AI to stay out
Blocks two of the four
Blocked
-
Your security service
Challenges every visit
Blocked
-
Pages that need scripts to appear
22 of 26
Blocked
-
Redirects, followed to the end
Land where they should
Holds
What we check and fix here · 4 of 21
- 01
We look at your site the way an AI does.
Not in a browser. We fetch it the way a crawler does and read what actually comes back, which is often much less than what you see.
› crawler-perspective fetch
- 02
We fix the instruction file the robots read.
One file on your server tells every AI whether it is allowed in. It has usually been there for years and nobody has read it since.
› robots.txt
- 03
We check that your security isn't blocking the AI along with the bots.
The service protecting your site from attacks often cannot tell an AI assistant from a scraper, and blocks both. From September 15 one of the biggest changes its defaults.
› WAF and bot-management rules
- 04
If the page only exists after JavaScript runs, we rebuild it so it arrives already written.
Some crawlers don't run JavaScript. If your text appears only after the page loads in a browser, to them the page is blank.
› server-side rendering
Free scan, no card required
Find out whether you're through this gate.
The free report scores all four gates and shows you the answers word for word, with the sources each one was built from.
On its way.
The report is generating now and lands in your inbox in a few minutes. Bring it to the walkthrough and we'll have read it too.
How visible is your site?
When buyers ask AI who to use, does your name come up or your competition's?
- Reading your site the way a model does
- Writing the questions your buyers ask
- Asking every major AI system
- Working out who gets named instead
This takes about five minutes. Keep this page open to watch it, or close it: the report reaches you by email either way.
Read your reportYour free scan is a starting point. The walkthrough is where it turns into a plan.
-
Who gets named instead
-
Where the answers came from
-
The answers, word for word
Checked across every major AI system
- ChatGPT
- Claude
- Gemini
- Perplexity
- AI Overviews
We'll email you the report. No card, no call required to get it.
Next gate: Understanding