We Checked 368 Small Business Sites. One in Four Are Partly Invisible to AI Crawlers.

Your website has two audiences now, and one of them cannot read JavaScript.

For a decade, the only rendering worry in SEO was Google. Can Googlebot see the content that only appears after your scripts run? That question is basically settled. Google renders. What changed is that a new class of crawler showed up, and it does not.

The crawlers behind the big AI assistants, OpenAI's GPTBot, Anthropic's ClaudeBot, and PerplexityBot, fetch your pages but do not execute the JavaScript on them. Google's renderer does. So anything your site paints in with JavaScript after the page loads is visible to Google and invisible to those AI crawlers. We wanted to know how often that actually hides something that matters, so we measured it.

What we did

We took 368 real small business websites and captured each one twice on the same pass. Once the way Google sees it, in a real headless Chromium browser with full JavaScript execution. Once the way a non-rendering AI crawler sees it, a plain fetch with a GPTBot user agent and no JavaScript. Then we compared the two and flagged any answer-critical element, body copy, headings, internal links, contact details, or a contact form, that was present for Google but missing from the raw fetch.

One framing note we hold to, because it is easy to overclaim: this measures what an AI company's own crawler retrieves on a direct fetch. It is not a claim that "ChatGPT cannot know you exist." Those assistants also pull from search indexes, licensed feeds, and citations. But the direct-crawl path is real, it is growing, and nobody had measured it for ordinary small business sites.

One in four sites hide something that matters

Across the 368 sites, 25 percent had at least one answer-critical element that Google could see and a non-rendering AI crawler could not. The single most common thing to go missing was the body text itself, the words that explain what the business does, hidden on 15.5 percent of sites. Internal links were next at 13.9 percent, which matters because links are how a crawler discovers the rest of your site.

And 14.7 percent hid their name, address, or phone number from a direct AI-crawler fetch. For a local business, that is the worst case: the page renders fine for a human and for Google, but a crawler asking "how do I contact this business" comes back with nothing.

The part that surprised us

The usual advice is that hosted site builders like Wix and Squarespace are bad for SEO, and that a custom build gives you control. For AI-crawler visibility, the data says the opposite.

The two platforms that server-render content by default, Wix and Webflow, were the safest by a wide margin. Wix sites lost content only 2.9 percent of the time, Webflow 6.7 percent. The worst results came from bespoke custom builds at 31.8 percent, the exact sites whose owners assume "full control" means safe. Squarespace (28.3 percent), WordPress (26.6 percent), and GoDaddy (20 percent) landed in between.

It is not builder versus custom. Squarespace is a builder and sits near the top of the risk list. The real predictor is whether the platform produces readable HTML before JavaScript runs. Wix and Webflow do that by default. Many custom React builds, and plenty of plugin-heavy WordPress and Squarespace sites, do not.

The full study, with the per-platform breakdown and the open dataset, is here.

How to check your own site in 30 seconds

You do not need a tool for the first pass. Open your homepage, right click, and choose View Page Source, the actual source, not Inspect. Then use Ctrl-F to search that source for three things: your phone number, your main heading, and your menu links.

If they are not in the source, they are being injected by JavaScript after the page loads. Google will run that JavaScript and see them. A non-rendering AI crawler will not.

The fix is one principle: if a piece of information matters for being found, it has to be in the HTML the server sends, not added later by JavaScript. That usually does not mean a rebuild. It means server-rendering your core content, keeping your contact details and navigation as real text and links in the source, and not letting a client-side widget be the only path to reach you.

See exactly what an AI crawler misses on your site

We built DeepAudit AI to render a page in a real browser and compare it to the raw HTML, so it flags precisely which elements are visible to Google but invisible to a non-rendering crawler. It is free, no signup, and it takes about 60 seconds.

Run the free audit on your own site. The report is yours to keep. If you would rather we read it with you, book a free 15-minute teardown and we will walk through what is exposed and how to fix it, live, no pitch.