When someone asks an AI assistant for a recommendation and your business doesn't appear, the assistant doesn't admit a gap — it names your competitors with full confidence. In our audits, the causes are remarkably consistent. Here are the five we see most, in order of how often they appear.

1. AI crawlers are blocked — often by accident

Security plugins, CDN bot protection and copy-pasted robots.txt templates routinely block GPTBot, PerplexityBot and friends. The result is absolute: a system that can't crawl you can't cite you. This is the first thing to check and usually the fastest to fix.

2. No structured data, so AI can't parse who you are

Schema.org markup (JSON-LD) is how you tell machines unambiguously: this is a business, here's its name, services, location, contacts. Without it, models reconstruct your identity from scattered HTML — and when they're unsure, they leave you out rather than risk a wrong answer.

3. Your content isn't citable

AI answers are assembled from clear, factual passages: definitions, numbers, direct answers to real questions. Pages that say "innovative solutions for your success" give an assistant nothing to quote. Pages that say "an implant costs $1,900–2,400 and takes two visits" get cited. Marketing fluff is invisible to the answer layer.

4. No entity footprint outside your own site

Models trust businesses they've seen in more than one place: industry directories, review platforms, maps, press, forums. If your brand exists only on your own domain, you're a claim with no witnesses. Consistent presence on the sources AI actually reads is what turns a name into an entity.

5. Inconsistent or outdated information across the web

Old addresses on a directory, different phone numbers on two platforms, a service list that changed three years ago — models see conflicting data and either pick the wrong version or skip you. An entity that contradicts itself is a low-trust entity.

The good news

Every one of these five is diagnosable and fixable. Start with the 5-minute self-test, add an llms.txt, and if you want the full picture — a GEO audit scores all five areas and hands your developer ready-to-paste fixes.