You've done the SEO work. Your site ranks on Google. But when someone asks ChatGPT or Perplexity to recommend a business like yours — you're nowhere to be found.
It's not because your product isn't good enough. It's because AI crawlers work completely differently from Google's bot, and most websites are accidentally blocking them without realising it.
At VisibAI, we've audited hundreds of websites for AI visibility. The same problems come up again and again. We call them AI Invisibility Traps — technical issues that silently cut you off from AI recommendations.
Here are all 8, what causes them, and exactly how to fix each one.
Trap 1: Cloudflare Bot Fight Mode 🛡️
Score penalty: -5 points
This is the single most damaging trap — and the most common one we find.
Cloudflare's "Bot Fight Mode" is a security feature that blocks automated bots from accessing your site. The problem: it doesn't distinguish between bad bots and AI crawlers like GPTBot (ChatGPT), ClaudeBot (Anthropic), or PerplexityBot.
When Bot Fight Mode is active, these crawlers get a 403 or 503 response. Your site is effectively a blank wall to every major AI platform.
The fix: In your Cloudflare dashboard, go to Security → Bots → set "AI Scrapers and Crawlers" to Allow. It's a single toggle. Takes 30 seconds. The impact on your AI visibility can be immediate.
Trap 2: JavaScript Rendering Trap ⚡
Score penalty: -4 points
AI crawlers cannot execute JavaScript. They fetch your raw HTML and that's all they see.
If your website loads its content dynamically via JavaScript — as many modern React, Vue, or Next.js sites do when not configured correctly — the crawler arrives at your homepage and finds almost nothing. No services listed. No about text. No pricing. Just an empty shell.
Google has trained its bot to render JavaScript. ChatGPT and Perplexity haven't. They rely on what's in the initial HTML response.
The fix: Enable server-side rendering (SSR) for your key pages. If you're on Next.js, make sure your landing page, about page, and services pages are rendered server-side or statically generated — not client-side only. Run curl https://yoursite.com in your terminal and check how much text comes back. If it's nearly empty, you have this problem.
Trap 3: Content Freshness Trap ⏰
Score penalty: -2 points
AI models treat undated content as stale. Research shows that pages with clear date signals receive significantly more citations from AI platforms than pages with no date metadata at all.
If your website has no Last-Modified header, no dateModified in your schema markup, and no visible dates on your pages, AI models have no idea when you last updated your content — and they'll deprioritise you in favour of sources they know are current.
The fix: Three things to add: (1) a Last-Modified HTTP header on your responses, (2) a dateModified field in your JSON-LD schema markup, and (3) visible publish/update dates on key pages like your blog posts, case studies, and service pages. If you update your homepage copy, update the date too.
Trap 4: Paywall / Login Wall Detected 🔒
Score penalty: -3 points
AI crawlers cannot log in. They cannot fill out forms, accept cookies, or authenticate. If your key content — your pricing page, your features page, your FAQ — is gated behind a login or a "sign up to continue" prompt, it is completely invisible to AI.
This is more common than you'd think. Some SaaS products gate their pricing. Some service businesses require contact form completion before showing their service list. Some e-commerce sites require login to browse.
The fix: Make your most important pages fully public. Pricing, features, FAQ, about, case studies — these should all be accessible without any authentication. The content AI needs to recommend you has to be freely readable by a bot.
Trap 5: Canonical URL Instability 🔗
Score penalty: -1 point
If your site is accessible at both https://yoursite.com and https://www.yoursite.com, or if you have redirect chains (http → https → www → non-www), AI agents can end up with an inconsistent picture of what your site is.
They may index two versions of the same page as separate entities, diluting your authority. Or they may follow a redirect chain and give up partway through.
The fix: Pick one canonical URL — either the www or non-www version — and redirect all others to it consistently. Set a <link rel="canonical"> tag on every page pointing to the definitive URL. Make sure your sitemap only contains the canonical versions. Check your redirect chain with a tool like curl -I https://yoursite.com and count the hops — ideally there should be zero or one.
Trap 6: PDF Content Trap 📄
Score penalty: -2 points
PDFs are difficult for AI crawlers. They can sometimes extract text from them, but structured HTML is always preferred. If a significant portion of your site's content lives in PDFs — menus, brochures, service catalogues, case studies — that content is largely invisible to AI.
This trap is especially common for restaurants (menus as PDFs), law firms (service brochures), and professional services businesses.
The fix: Convert your most important PDF content to HTML pages. A PDF menu becomes a /menu page. A PDF brochure becomes a /services page. A PDF case study becomes a /case-studies/client-name page. Not only does this fix the AI visibility problem — it also improves your Google rankings.
Trap 7: Entity Naming Inconsistency 🏷️
Score penalty: -1 point
AI models build a mental model of your business as an entity — a named thing with a consistent identity. If your brand name appears differently across your own site, it creates confusion about who you actually are.
For example: your page title says "Acme Ltd", your og:site_name says "Acme", your schema says "Acme Digital Solutions", and your footer says "© Acme Co." These all look like different entities to an AI trying to build a coherent picture of your business.
The fix: Audit every place your brand name appears — page title, og:site_name, JSON-LD Organization schema, footer, meta description — and make them all use the exact same name. Pick one form and apply it everywhere consistently.
Trap 8: Content Hidden in Interactions 📦
Score penalty: -1 point
Tabs, accordions, modals, and "read more" toggles are great for user experience. They're terrible for AI crawlers.
When content is hidden behind a click interaction and uses display: none in the default HTML, AI crawlers may not register it exists. Your FAQ section hidden inside an accordion? Invisible. Your service details behind a tab? Invisible. Your pricing breakdown inside a modal? Invisible.
The fix: For any content that's critical for AI recommendations — FAQs, service descriptions, pricing explanations — make sure it exists in the visible default HTML, even if it's visually collapsed. The text needs to be present in the DOM from the server-side render. Alternatively, move key content out of interactive components entirely and onto dedicated pages.
How Many Traps Is Your Site Triggering?
Most businesses we audit are triggering at least two or three of these traps without knowing it. A few are triggering five or six — effectively invisible to every major AI platform despite having a well-designed, content-rich website.
The frustrating thing is that most of these fixes are straightforward. The Cloudflare toggle takes 30 seconds. Adding date metadata takes an hour. Enabling SSR on a Next.js site is often a configuration change.
The hard part is knowing which traps you're in.
That's what VisibAI is for. Run a free audit at getvisibai.com and we'll check all 8 traps automatically, alongside 52+ AI queries across ChatGPT, Perplexity, Claude, Gemini, and Mistral. You'll get a score out of 100, a full fix list, and downloadable fix files — ready to deploy.