When someone asks ChatGPT 'who's the best [your service] in [your city]', it doesn't run a live Google search. It draws on a broad picture of the web assembled during training and retrieval — synthesising what it knows about businesses in your category to produce a confident recommendation.
That picture is built from multiple signals. Understanding what those signals are — and which ones your business is currently weak on — is the first step to improving your AI visibility.
How AI search works
What ChatGPT looks at when recommending a business
ChatGPT doesn't rank websites like Google does. It builds a picture of your business from multiple signals across the web — and recommends the businesses it considers most credible and clearly described. Here's exactly what it's looking at.
How AI tools evaluate businesses
Structured data and schema markup
Structured data is machine-readable code added to your website that explicitly tells AI systems what your business is, what it does, where it operates, and how to categorise it. Without it, AI tools have to infer your offering from the words on your pages — and they often get it wrong or lack the confidence to include you.
The most valuable schemas for AI visibility are Organisation (establishes your identity), LocalBusiness (establishes your location and service area), Product or Service (describes what you offer), and FAQPage (signals that you directly answer questions in your sector).
Most business websites have no structured data at all. Adding it is one of the highest-impact, most direct interventions available.
The most valuable schemas for AI visibility are Organisation (establishes your identity), LocalBusiness (establishes your location and service area), Product or Service (describes what you offer), and FAQPage (signals that you directly answer questions in your sector).
Most business websites have no structured data at all. Adding it is one of the highest-impact, most direct interventions available.
Third-party mentions and citations
AI tools treat what others say about your business as significantly more credible than what you say about yourself. This mirrors how humans evaluate recommendations — a trusted friend's endorsement carries more weight than a brand's own advertising.
In practice, this means AI systems weight mentions on trusted external sites heavily: industry directories, professional associations, review platforms, news coverage, analyst reports, and roundup articles. The more authoritative the source, the more weight it carries.
Businesses with rich third-party citation profiles — even if their own websites are modest — often outperform businesses with polished sites but weak external presence.
In practice, this means AI systems weight mentions on trusted external sites heavily: industry directories, professional associations, review platforms, news coverage, analyst reports, and roundup articles. The more authoritative the source, the more weight it carries.
Businesses with rich third-party citation profiles — even if their own websites are modest — often outperform businesses with polished sites but weak external presence.
Key points
What you say about yourself matters less than what others say
AI tools discount self-description in favour of third-party signals. A mention in a reputable trade publication is worth more than five pages of marketing copy.
Structured data is your direct communication channel with AI
Schema markup is the one place where you can speak directly to AI systems in their native language. Businesses without it rely on AI to guess — and guesses lead to omission.
Consistency is a trust signal
When AI tools find conflicting information about your business across different sources, they lose confidence. Confident, consistent signals across the web lead to confident recommendations.
Content that answers specific questions
AI tools are fundamentally question-answering systems. They're trained to identify which content reliably answers questions — and they favour businesses whose content does this well.
This doesn't mean writing for keywords. It means having pages and articles that genuinely address the questions your prospective customers ask: what you do, who you help, how you compare to alternatives, what the process looks like, what it costs.
Content that's specific, accurate, and genuinely useful gets used as a source. Generic content that could describe any business in your category gets ignored.
This doesn't mean writing for keywords. It means having pages and articles that genuinely address the questions your prospective customers ask: what you do, who you help, how you compare to alternatives, what the process looks like, what it costs.
Content that's specific, accurate, and genuinely useful gets used as a source. Generic content that could describe any business in your category gets ignored.
Entity recognition and brand consistency
AI systems build 'entities' — internal representations of real-world things, including businesses. For your business to have a strong entity, it needs to be consistently described across every source the AI can see.
Your business name, description, location, founding year, and core offering should be the same on your website, Google Business Profile, Companies House listing, LinkedIn, Trustpilot, industry directories, and anywhere else you appear. Inconsistencies fragment your entity and reduce the AI's confidence in recommending you.
This is often the most straightforward issue to fix — and one of the most commonly overlooked.
Your business name, description, location, founding year, and core offering should be the same on your website, Google Business Profile, Companies House listing, LinkedIn, Trustpilot, industry directories, and anywhere else you appear. Inconsistencies fragment your entity and reduce the AI's confidence in recommending you.
This is often the most straightforward issue to fix — and one of the most commonly overlooked.
Frequently asked questions
Does having lots of Google reviews help with ChatGPT visibility?
Yes, indirectly. Google reviews are one form of third-party credibility signal that AI tools can access. A strong review profile on Google, Trustpilot, or sector-specific review platforms contributes to the overall picture of your business as a trusted, established option. Reviews alone aren't sufficient, but they're a meaningful part of the picture.
Does social media presence affect ChatGPT recommendations?
To a limited extent. Social media content is generally weighted lower than editorial or directory content, but a strong, consistent social presence can contribute to entity recognition — particularly LinkedIn for B2B businesses. The more important factor is being mentioned on sites that AI tools consider authoritative sources, rather than the volume of your own social posts.
My website is well-designed and informative — why don't I appear in AI results?
A good website is necessary but not sufficient for AI visibility. The factors AI tools weight most heavily — structured data, third-party citations, entity consistency — are largely invisible on a well-designed site. A polished homepage with no schema markup and no external mentions will consistently underperform a simpler site that has strong signals in all three areas.
How is Perplexity different from ChatGPT in terms of what it looks at?
Perplexity places greater emphasis on real-time web retrieval — it actively searches the web for each query and synthesises the results. This means being findable via Google (ranked pages, structured data) matters more for Perplexity than for ChatGPT, which relies more on its training data. That said, the core signals — credibility, clear description, third-party presence — are valuable across both platforms.