Less than 3% of businesses appear when AI systems are asked about their category. That is not a rounding error or a temporary glitch — it is a structural feature of how AI search works. When someone asks ChatGPT, Perplexity, or Gemini to recommend an accountant, a software agency, or a local plumber, the AI draws from a model it has built of the world. If your business was not built into that model correctly, it does not appear. Full stop.
The 97% figure surprises business owners, because most assume visibility is automatic. They have a website, they rank somewhere on Google, they have social profiles — surely AI systems can find them? The answer, in most cases, is no. AI search does not work like Google. The signals it relies on are different, the logic is different, and the businesses that appear consistently have done specific things to earn that place.
Why AI search works differently from Google
Google uses PageRank — a system that scores pages based on the volume and quality of links pointing to them, then matches those pages to search queries. It is a retrieval system: a user types a query, Google fetches the most relevant pages, and ranks them in order.
AI search works on an entirely different principle. It does not retrieve pages — it builds a model of the world from everything it has read, then draws on that model to answer questions. When an AI system recommends a business, it is not fetching a page in real time. It is recalling what it knows about that business from training data and live retrieval: who they are, what they do, who trusts them, and whether the evidence for that trust is consistent and corroborated.
This means the signals that drive AI search visibility are fundamentally different from SEO signals. There is no PageRank equivalent. There are no title-tag keywords. What matters is knowledge — does the AI have a coherent, confident model of your business? — citations — have credible third-party sources mentioned you in context? — and entity clarity — is your business consistently described across the web in a way AI systems can read and trust?
A business can rank on page one of Google and be entirely absent from AI answers. A modest website with strong citations and clear entity signals can appear prominently. The two systems do not overlap the way most people assume.
What the 3% do differently
The businesses that consistently appear in AI search have not stumbled into visibility. They have — intentionally or not — satisfied the signals AI systems rely on. Four patterns distinguish them from the invisible majority.
Their pages use clean HTML with valid schema markup. Core information — what the business does, who it serves, where it operates — is accessible in plain text without JavaScript execution. AI crawlers can read it without friction.
They appear in sector publications, trade directories, and trusted review platforms — not just on their own website. Each citation is a data point that corroborates their existence and category in the AI’s model.
Their business description — what they do, who they serve, what category they belong to — is identical across their website, Google Business Profile, LinkedIn, and directory listings. AI systems can build a confident model of the entity because it never contradicts itself.
Their robots.txt explicitly permits GPTBot, ClaudeBot, and PerplexityBot. They have not accidentally blocked the crawlers that feed AI systems — a mistake that affects a surprising number of otherwise well-built sites.
The 5 biggest reasons businesses stay invisible
Conversely, the 97% who are invisible share a predictable set of problems. These are not obscure edge cases — they are the most common failure modes Rankara sees when auditing businesses across every sector.
Missing or ambiguous entity
AI systems think in entities — clearly defined real-world objects. If your business description is vague (“we help companies grow”), inconsistent across platforms, or missing structured data, the AI cannot build a confident model of who you are. Ambiguous entities are excluded — not out of hostility, but because the AI cannot describe you with confidence.
No third-party citations
Self-published content — your website, your blog, your social posts — carries limited weight in AI systems. What moves the needle is corroboration from independent sources: trade publications, sector directories, review platforms, press coverage. Without citations, the AI has only your word for your own existence, which is not enough to confidently recommend you.
Blocked AI crawlers
Many businesses inadvertently block GPTBot, ClaudeBot, or PerplexityBot in their robots.txt — often as an accidental side effect of blocking all bots by default. If an AI crawler cannot read your site, it cannot extract the content that builds your model. This is a hard technical block that no amount of off-site citation work will fully overcome.
Thin or non-quotable content
AI systems favour content they can extract as clear, authoritative statements. Generic marketing copy (“we deliver exceptional results for our valued clients”) provides nothing quotable. Thin pages with no substantive claims about expertise, methodology, or outcomes give the AI nothing useful to work with. The result: even if the AI knows your business exists, it has nothing worth surfacing about you.
No mention ecosystem
A business that exists only on its own website and a handful of social profiles has a thin mention ecosystem. AI systems corroborate recommendations against the breadth and quality of a business’s web footprint. No press coverage, no sector directory presence, no review platform entries, no knowledge graph entry — each gap is a missing signal that makes the AI less confident about recommending you.
How to find out where you stand
The first step is knowing your current AI search visibility score. Rankara’s free tool checks your domain against the core signals AI systems use — entity clarity, citation presence, crawler access, and content structure — and gives you a score in under a minute. No account required.
If your score is low, or if you want to understand exactly what is holding you back across all nine dimensions of AI visibility, the Rankara AI Visibility Audit goes deeper. It runs 52 checks across your entire digital presence — website, citations, schema markup, robots.txt, mention ecosystem, and more — and delivers a prioritised fix roadmap you can act on immediately.
The businesses that move into the visible 3% do not do so by accident. They audit, they fix the specific gaps, and they build the signals AI systems need. The gap between the visible minority and the invisible majority will only widen as AI search handles more commercial queries in 2026 and beyond. The earlier you address it, the smaller the work required.
For a deeper look at the specific factors AI systems use to decide who to recommend, see: The 5 AI search ranking factors in 2026 (that have nothing to do with SEO).
Find out if you’re in the 3%
Get your Rankara AI Visibility Audit for €99. A full 9-section report scoring your business across every AI search signal — with a prioritised fix roadmap so you know exactly what to do next.
Get Your AI Visibility Audit →