Your website structure: why AI can read some sites and not others
Some websites are unreadable to AI systems, and the businesses behind them usually have no idea, because the site looks fine in a browser. Readability failures are mechanical: a blocked crawler, content that only appears when JavaScript runs, text trapped inside images or PDFs. Before any other visibility work makes sense, this one has to pass, and you can test most of it yourself while reading this post.
The mechanics of being invisible
A browser and an AI crawler read your site differently. The browser runs your scripts, loads your images, and shows a human the result; many crawlers read the raw page text and move on. If your content only exists after scripts run, inside images, or behind a crawler block, the site a machine sees can be nearly empty while the site you see looks great.
The three-minute self-test, in the order to run it
Check one: view the source. Open your site, view the page source (right-click, “View Page Source”), and search for a sentence you know is on the page. If your services and city appear in the raw source, machines can read them; if the source is a wall of script tags with none of your actual words, you’ve found the problem.
Check two: the robots question. Visit yoursite.com/robots.txt and look for lines that disallow everything or name AI crawlers. You might have chosen that; more often a security plugin or CDN default chose it for you, and the full log-checking walkthrough lives in that post.
Check three: the phone test. Load your site on a phone on cellular. If it takes ages or key content never appears without tapping something, machines on a time budget are seeing the before state.
What AI-readable structure looks like
Plain text in real HTML, headings that say what the section answers, and the important facts (what you do, where, licensing, availability) stated in sentences rather than implied by design. One annotated example, labeled as illustration: a service page whose H2 reads “Water heater replacement in Fort Collins, usually same day” followed by three plain sentences is liftable by a machine word for word. The same information as white text over a hero photo, rendered by script, is invisible twice over. The construction manual for pages like this is its own walkthrough.
The builds that cause trouble
Honest severity, no fearmongering: JavaScript-only rendering (common in modern site builders; some crawlers run scripts, many don’t, so you’re gambling on which reads you). Menus, price lists, and service details as PDFs or images of text, which is the restaurant industry’s signature own goal. And blanket crawler blocks from security tooling, the one failure that costs everything and announces nothing. None of these is fatal to fix, and most are one afternoon with the diagnostic flow.
Where this sits among the signals
Readability is checked in every free scan because it gates the rest: reviews and mentions can’t rescue a site machines can’t parse. Our opinion, from watching this industry oversell: readability is also the most over-fixed signal once it passes. If your source-view check shows your words, stop optimizing here and go feed the sources that actually decide mentions; rebuilding a readable site for marginal gains is where budgets go to die.