Can AI Read Your Website? How ChatGPT and Google's AI See Your Business

Can AI read your website? Only what shows up in the page's raw text response, not what a person sees on screen. If your content is written in by JavaScript after the page loads, or your key facts sit only inside an image, an AI assistant can receive a page that is almost entirely blank. It cannot recommend a business it cannot read.

This matters because ChatGPT, Perplexity, and Google's AI features all work the same way at the fetching stage: something requests your page, gets back a response, and reads the text in that response. No assistant looks at your site the way a customer does, with a rendered screen and working eyes. It reads a document. What is and is not in that document decides whether you exist to it at all.

How a crawler actually reads a page

When you open your website in a browser, a lot happens before you see anything: the browser fetches the HTML, then fetches and runs any JavaScript, then builds the page you actually look at. That finished, built page is what a person sees.

A crawler or an AI assistant fetching a page for its own use typically requests the same public URL, but many of them stop at the first step. They read the raw HTML response the server sends back, before any JavaScript has had a chance to run. If your text, headings, and business details are already written into that raw response, the crawler gets them. If they only appear after a script runs in a browser, the crawler often gets nothing.

Some AI systems and search crawlers have gotten better at running JavaScript before reading a page, but it is slower, more expensive, and not guaranteed. Treat "my page depends on JavaScript to show its content" as a real risk, not a solved problem.

Text trapped in images

A photo of your menu, a graphic listing your hours, a logo with your tagline baked into the design: none of that text is machine-readable text. An assistant reading your page sees a reference to an image file, not the words inside it. Unless that same information also exists as real text somewhere on the page (or in descriptive alt text, which helps but is not a substitute for the words being on the page), it is invisible to both search engines and AI assistants.

This is a common trap for businesses that design a nice-looking hours-and-location graphic for the website and never type the same details anywhere as plain text.

Slow pages and the mobile version

Crawlers mostly evaluate the mobile version of your site, not desktop, because that is what Google indexes primarily now and what most real visitors use. If your mobile page is slow, cuts off content, or hides sections behind extra taps that require JavaScript to reveal, that stripped-down version is what gets judged, both for ranking and for whether an assistant can pull anything useful from it.

How to check: ask "is my website mobile friendly" and answer it with Google's own free checks rather than guessing: the Mobile-Friendly Test and the Rich Results Test, both run against your live URL and show you what Google's crawler actually renders and finds. If your content is missing or cut off there, treat it as missing everywhere.

Robots rules and AI crawler access

Your robots.txt file can allow or block specific crawlers by name. Some AI companies operate their own crawler and respect a disallow rule the same way Google does. If you, a past developer, or a website builder default has blocked one of these crawlers, that assistant cannot fetch your pages at all, and cannot mention your business based on your site's own content no matter how good that content is.

How to check: visit yourdomain.com/robots.txt directly in your browser and read through the user-agent rules. A blanket Disallow: / under a specific crawler's name blocks that crawler entirely.

Should you block AI crawlers? Blocking one removes you from what that assistant can say about you directly. For most small and local businesses trying to be found, that is a cost with no offsetting benefit. It is a deliberate decision to make with your own reasons, not something to leave to a default setting you never reviewed.

Business facts that live only in a widget or an image

Your hours, address, phone number, and services are exactly the facts an AI assistant needs to answer a customer's question about you. If those facts only exist inside a map widget, a booking plugin, or a graphic, and never as plain text on the page itself, an assistant reading your site cannot confirm any of it. It will either say nothing about you or fall back on outdated or third-party information instead.

Put the essentials in real text, even if a widget also displays them: your name, address, phone number, hours, and a plain description of what you do and where you do it. This is a small amount of writing that closes a real gap.

Consistency across the web

Assistants and search engines also cross-check what your site says against what directories, review sites, and your Google Business Profile say. If your address or hours differ across those sources, that inconsistency is itself a signal that lowers confidence in all of them, even when your own website is perfectly readable.

How to check it yourself

• View source, search your own text. Ctrl+U on your homepage, then Ctrl+F for your business name or a sentence from your main heading. If it is not in the raw source, it was likely added by JavaScript.
• Disable JavaScript and reload. Most browsers let you turn JavaScript off for a page (in Chrome, through the site settings or developer tools). Reload and see what is left. Whatever disappears is invisible to a crawler that does not execute scripts.
• Run Google's Mobile-Friendly Test and Rich Results Test. Both are free, both show you the rendered HTML Google actually captured from your live page, and both will surface missing content directly.
• Search for your own text in quotes. Take a specific sentence from your site, search it in quotation marks in Google. If Google has never indexed that exact text, it likely never read it, which means an AI assistant drawing on the same index is working from the same gap.

A site Google cannot read is, for the same underlying reason, a site AI assistants cannot read either. They are pulling from overlapping sources and hitting the same rendering problem. That is also why the fixes overlap: solving the crawlability issue tends to help both at once, which is worth keeping in mind if you are also working through why a site is not showing up on Google.

FAQ

How do I check what a crawler sees on my site?

View the page source (Ctrl+U) and search for your own text, or use Google's free Mobile-Friendly Test and Rich Results Test, which show the rendered HTML Google captured from your live URL. If your key text is missing from either, that is what a crawler sees: nothing there.

Does JavaScript hurt SEO and AI visibility?

JavaScript itself is not the problem. The problem is content that only appears after JavaScript runs, with nothing equivalent present in the raw HTML response. Some crawlers run scripts before reading a page, many do not, or do so unreliably. If your key text depends entirely on a script to appear, treat it as a real visibility risk.

Should I block AI crawlers?

Blocking a named AI crawler in robots.txt removes you from anything that assistant can say about you based on your own site. For most small and local businesses trying to be found and recommended, that is a real cost with little upside, so it is worth a deliberate decision rather than an unreviewed default.

Can AI read text in images?

No, not as page text. An assistant reading your page sees that an image exists, not the words inside it. Put essential facts, like your hours, address, and services, into real text on the page as well, even if you also show them in a graphic or widget.