JB Aten  • July 27, 2026

What AI Engines Actually Read on Your Website (and What They Skip): A GEO Reality Check

AI engines like ChatGPT don't read your website the way people do. Here's what they actually pull from your pages — and what they ignore completely.

Here’s something most business owners don’t know: when ChatGPT or Perplexity “reads” your website, it usually spends less than a second there — and it never sees your hero image, your slider, your video, or that beautiful animation you paid for. It grabs text, and only text it can reach quickly. Everything else might as well not exist.

That matters more every month. Bain reported in 2025 that about 80% of consumers now rely on AI-generated summaries for at least some of their searches, and roughly 60% of searches end without a single click to a website. Gartner projects traditional search engine volume will drop 25% by 2026 as people shift to AI assistants. If AI engines can’t read your site, you’re invisible to a growing share of your customers — even if your Google rankings look fine.

So let’s get specific about what these engines actually consume.

What AI engines read

When an AI engine (or the crawler feeding it) hits your page, it prioritizes:

  • Plain text in the HTML. Paragraphs, headings, lists. If the words are in the page source, they get read.
  • Headings that state facts. “Emergency plumbing in Prescott, AZ — 24/7 response” tells an AI exactly who you serve. “Welcome to Our Site” tells it nothing.
  • The first few hundred words. Engines summarize. Content buried at the bottom of a long page gets less weight than what’s up top.
  • Specific, quotable statements. Numbers, years in business, service areas, prices, credentials. AI answers are built from concrete claims, and engines prefer sources that supply them. This isn’t a guess — the Princeton GEO study presented at ACM KDD 2024 found that adding statistics, quotations, and cited sources to a page increased its visibility in AI-generated answers by up to 40%. Vague pages lost ground; specific pages gained it.
  • Structured data. Schema markup (a behind-the-scenes label system that tags your business name, address, hours, and reviews) gives engines pre-digested facts they can trust.
  • Your llms.txt and About page. Engines look for a clear, plain statement of who you are, what you do, and where.

What AI engines skip

  • Images and video. A photo of your storefront with no descriptive text is invisible. Text baked into images (menus as JPEGs are the classic offender) is unreadable.
  • Content loaded by JavaScript. If your reviews, service list, or pricing only appear after scripts run, many AI crawlers never see them. They read the raw HTML and move on.
  • Sliders, tabs, and accordions. Content hidden behind clicks is often deprioritized or missed entirely.
  • Marketing fluff. “Best-in-class solutions for all your needs” contains zero extractable facts. Engines don’t cite adjectives; they cite claims.
  • PDFs buried in your site. Some engines read them, most don’t bother. Anything important in a PDF should also live on a normal page.

What to do about it today

You don’t need a redesign. You need your facts in plain text. Here’s a checklist you can work through this afternoon:

  1. Write one “answer paragraph” per key page. Two to three sentences near the top stating exactly what you do, where, for whom, and since when. Example: “Smith HVAC has repaired and installed air conditioning systems in Prescott and Prescott Valley since 2009. We’re licensed, bonded, and offer same-day service.”
  2. Turn image-only information into text. If your services, menu, or pricing exist only as a graphic, retype them onto the page.
  3. Add three specific numbers to your homepage. Years in business, jobs completed, review count, response time — anything true and concrete. The Princeton KDD research showed statistics are one of the highest-impact additions you can make.
  4. Replace vague headings. “Our Services” becomes “HVAC Repair, Installation & Maintenance in Prescott, AZ.”
  5. Check your site with JavaScript off. In Chrome, disable JavaScript in settings and reload your page. Whatever disappears is what many AI engines never see.
  6. Add basic schema markup. At minimum, LocalBusiness schema with your name, address, phone, hours, and service area. Your web person can add this in under an hour.

The bottom line

AI engines are literal-minded readers with no patience. They reward websites that state facts plainly, in text, near the top of the page — and they skip everything designed purely to impress human eyeballs. The businesses winning AI citations right now aren’t the ones with the flashiest sites. They’re the ones whose pages read like a well-organized fact sheet.

Want to know what AI engines currently say about your business? Get your free AI Visibility Score at rankforward.ai/score — RankForward will show you exactly how ChatGPT and Perplexity see you today, and what to fix first.

Is AI recommending your business?

Find out free. We ask ChatGPT and Perplexity about your business and send your score within 24 hours.

Get Your Free AI Score →