Here is a problem that produces some of the strangest audit results I see. A business has a modern, expensive, good-looking website. Visitors love it. And Google barely ranks it for anything, while AI assistants act like it does not exist. The site looks perfect and performs like it is invisible.
The cause, surprisingly often, is how the site is built. Many modern websites do not actually send your pages to the browser. They send an almost empty shell plus a bundle of JavaScript, a program that then runs on the visitor’s device and builds the page on the spot. For a human with a modern phone, the result looks identical. For a machine that fetches your page and reads what came back, the difference is enormous: one version contains your content, the other contains a shell and a script.
Who reads your site without running the program
Google can run JavaScript, with caveats: rendering costs it real resources, so JavaScript-heavy pages can be processed in a delayed second wave, and if anything in the setup fails quietly, Google indexes the empty shell without telling anyone. It works until it does not, and when it does not, nothing visibly breaks.
The bigger problem in 2026 is everyone other than Google. Many AI crawlers, the ones feeding ChatGPT, Claude and Perplexity, fetch pages and read what comes back without executing anything. If your content only exists after a program runs, these systems receive a page that says essentially nothing. At exactly the moment AI assistants became a real source of customers, a whole category of websites became unreadable to them.
The two-minute check
You do not need a developer for a first diagnosis. Open your most important page, press Ctrl+U (Cmd+Option+U on a Mac) to view the page source, and search for a sentence you can see on the page. Found it in the source? Your content ships as real text. Not found, amid a wall of code? Your page is assembled by script, and machines that do not run scripts see none of it.
A second check: your page might be indexed but hollow. Search Google for site:yoursite.com to see what is indexed, and if pages are missing entirely, the indexing guide walks through the other causes.
The fix is a decision, not a patch
None of this means JavaScript is bad; it makes sites pleasant and interactive. The question is only where the page gets built. Built in the visitor’s browser, machines get a shell. Built ahead of time, so the server sends finished pages with all the text already in them, machines get everything, and as a bonus the site gets faster for humans too, which feeds directly into Core Web Vitals.
Modern tooling makes the second option entirely practical; every site I ship works this way. Pages arrive as complete, readable text for Google, for AI crawlers, and for visitors, with JavaScript reserved for the small touches that genuinely need it. If you are commissioning a site, put the question to your builder in exactly these words: “will the content be in the page before any script runs?” The answer tells you a lot about what you are buying, and I have written about what else to ask when paying for a site.
If your current site fails the view-source test, that is fixable, sometimes without a rebuild. It is one of the first things I check in the free audit, because when this is broken, no amount of content work on top of it pays off. And if a rebuild is the honest answer, that is the core of what Studio Squalli does.