Someone in your city opens ChatGPT and types: “Who are the best [your service] providers in [your city]?” Three competitors get named. You don’t.
That moment is happening right now — thousands of times a day across every service category. And for most businesses, the absence is invisible. There’s no notification, no data point, no monthly report flagging it. The customers who would have called simply don’t know you exist.
The question every business owner asks when they discover this: how does the AI decide who to recommend? The answer is more systematic — and more fixable — than most people assume.
How AI systems build their recommendations
When you ask ChatGPT, Perplexity, or Gemini to recommend a business, it does not search the web in real time and rank results by authority score. It synthesises an answer from its understanding of which businesses exist in a category, what they do, and how credible they appear based on the information it has been trained on or can retrieve.
That understanding is built from thousands of data points: your website, directory listings, press mentions, social profiles, structured data, third-party articles, review platforms, and any other text on the web that associates your business name with a service category. AI systems aggregate this picture and decide: is this business clearly identifiable? Is it credible? Can I describe it accurately?
Businesses that score well on those three questions get recommended. Businesses that don’t — no matter how good they actually are — get excluded. The selection is not about quality of service. It’s about clarity of signal.
The 3 factors that determine who gets recommended
These are the signals that consistently separate businesses that appear in AI recommendations from those that don’t.
Entity clarity — does the AI know what you are?
An “entity” in AI terms is a coherent, identifiable thing in the world: a person, a company, a place. For AI systems to recommend your business, they must have a clear, consistent model of your entity — your name, your services, your location, and the type of client you serve.
If your website uses vague language (“full-service solutions partner”), if your description differs between your website, LinkedIn, Google Business Profile, and directory listings, or if you have no structured data explicitly identifying your business type — AI systems cannot form a reliable model of who you are. Ambiguous entities are not recommended.
LocalBusiness or Organization schema on your homepage, a consistent one-paragraph business description used across all platforms, and service pages that explicitly state what you do, for whom, and in which locations.Citation authority — are credible sources talking about you?
AI systems trust businesses that are mentioned in authoritative external sources. Not backlinks in the SEO sense — actual text that associates your business name with your service category in a credible editorial context.
A mention in a sector trade publication, a listing in a curated professional directory, a case study cited in a third-party article, an award recognition in an industry newsletter — these citations build AI credibility. They tell the AI system that your business is real, recognised, and relevant to its category. A business that only exists on its own website has no external corroboration. AI systems treat it as unverified.
Content structure — can AI extract and cite you?
AI systems favour content that is easy to parse and quote directly. That means pages with clear headings, direct answers to specific questions, and factual claims that stand on their own without requiring the full context of a page to make sense.
A 1,500-word brand narrative with no subheadings is hard to cite. A service page with a section titled “What does [Company] offer in [City]?” followed by a direct, factual answer is an ideal citation target. The AI can lift that passage, verify it against other sources, and include it in a recommendation with confidence.
Why this is different from paid ads and traditional SEO
Paid ads and traditional SEO are transactional: you pay or you optimise, and you appear. The mechanism is direct. AI recommendations do not work this way — and that is both the challenge and the opportunity.
You cannot buy your way into an AI recommendation. There is no bid, no sponsored slot, no paid placement in ChatGPT’s response to “recommend me a financial adviser in Amsterdam.” AI systems do not accept advertising for their organic recommendations — they synthesise them from what they understand to be true about the world.
Traditional SEO, meanwhile, optimises for Google’s ranking algorithm — a system built around link authority, keyword relevance, and page architecture. Those signals matter far less to AI systems. A business with strong Google rankings but weak entity clarity and no third-party citations will consistently be outrecommended by a competitor with a clearer AI visibility profile, even if that competitor ranks lower on Google.
The implication is significant: AI visibility requires a different strategy. Not better ads. Not more keyword optimisation. A systematic investment in entity clarity, citation breadth, and content structure — the signals AI systems actually use.
The 3 most common reasons businesses get excluded
In Rankara’s audits, the same three patterns account for most exclusions from AI recommendations.
1. No machine-readable identity
The business has no structured data on its website — no schema markup identifying its business type, location, services, or operating hours. AI systems cannot reliably identify the entity without this signal. The website may be well-designed and well-written, but to an AI system scanning for structured information, it is opaque. Fix: add LocalBusiness or Organization schema to your homepage and key service pages.
2. No external corroboration
The business exists primarily on its own website and social channels, with minimal presence in third-party editorial sources. AI systems treat self-declared credentials with scepticism — they look for independent corroboration that a business is who it says it is. Without press mentions, directory listings, or third-party citations, there is nothing to corroborate. Fix: prioritise a targeted list of high-quality directory and publication placements in your sector.
3. Inconsistent or vague descriptions
The business describes itself differently across platforms — or uses language so generic that the AI cannot confidently categorise it. “We help businesses grow” tells an AI system almost nothing about what you do or who you serve. Inconsistency across your website, Google Business Profile, LinkedIn, and directories creates conflicting signals. Fix: write a clear, specific, one-paragraph business description and deploy it consistently everywhere your business appears.
A 5-minute check any business owner can do right now
You don’t need a full audit to get a first read on where you stand. Run through these five checks in the next five minutes:
- Ask ChatGPT or Perplexity directly.Type: “Who are the best [your service category] in [your city]?” Do you appear? Note which competitors do — they are your AI visibility benchmark.
- Check your structured data. Go to search.google.com/test/rich-results and enter your homepage URL. Does Google find valid schema markup? If the test returns no structured data, AI systems have no machine-readable identity for your business.
- Google your own business name. Do your Google Business Profile, LinkedIn, and website all describe you in the same terms? If the descriptions conflict or differ substantially, you have an entity consistency problem.
- Count your third-party citations. Search for your business name in quotes. Beyond your own website and social profiles, how many independent editorial sources mention you? If fewer than five, your citation footprint is below the threshold AI systems require for confident recommendations.
- Read your homepage as an AI would.Does the first paragraph state clearly what you do, for whom, and where? Or does it open with a brand claim (“we’re passionate about excellence”)? AI systems need factual, specific language — not marketing copy.
If you found gaps in more than two of those five checks, your AI visibility has meaningful room for improvement. The good news: all of them are fixable with the right prioritisation.
For a complete picture — including a scored assessment across all nine AI visibility signals, a competitor benchmark, and a 90-day fix roadmap — see what a full audit covers in our sample AI Visibility Report.
Find out where you stand in AI search
Get your free AI Visibility Score in 30 seconds. See exactly what ChatGPT, Perplexity, and Gemini find when they evaluate your business — and what’s keeping you off the recommendation list. Full 9-section audit for €99.
Get My Free AI Visibility Score →