When you ask ChatGPT to recommend a hotel in Bangkok, or a wealth manager in New York, or a rehab facility in Cyprus, it doesn't consult Google's rankings. It doesn't pull a fresh list of backlinks. It reasons from training data—patterns of which entities appear together, which ones hold authority signals, and which ones are talked about in contexts that feel trustworthy. This is a fundamentally different process from traditional search optimisation, and if your business isn't visible in these recommendations, you're losing customers who've already decided to ask an AI.
The shift matters because ChatGPT and Claude and Gemini are now the first interface many people use when they're looking for a specific service. They're not running a search engine. They're generating text. And the businesses they choose to mention come from a model trained on the internet—which means your visibility depends not on links or page speed, but on how well your entity is represented across trustworthy sources that the model learned from.
Entity Recognition and Coherence
ChatGPT doesn't have a database of businesses. It has learned patterns about how language works, and that includes learning which names and entities are real, which ones matter, and which ones appear in clusters that signal authority. When you ask it to recommend a business, it's essentially saying: "I've learned that this entity name, when mentioned in training data, appears alongside signals of trust and credibility."
This is why consistent naming across sources matters more than it ever did in Google SEO. If your business is listed as "Wellness Retreat Cyprus" on your site, "Wellness Retreat" on LinkedIn, "The Wellness Retreat" on Google Business, and "Wellness Retreat Medical Center" in healthcare directories, you're fragmenting your entity. The model sees these as separate things. Consolidate your name, your description, and your core services into a single, consistent entity across every platform where you exist. This coherence is what signals to the model that you are a real, unified thing worth remembering.
The second part of entity coherence is context. ChatGPT learns relationships between entities—what industries they belong to, who their peers are, what kind of work they do. If your site and your citations all position you in one vertical (say, luxury hospitality), but you're also mentioned in articles about geriatric care, the model sees a weak or confused entity. Consistency in your vertical strengthens how ChatGPT learns to classify and recommend you.
Training Data Freshness and Trustworthiness
ChatGPT's knowledge cutoff means it doesn't browse the live web for recommendations. It draws from patterns in data it learned from during training. That training data includes news articles, published research, business directories, and what might be called the "respectable internet"—sources that are themselves authoritative.
This means your visibility in ChatGPT recommendations is heavily weighted towards mentions in published sources that the model treats as trustworthy. A mention in a major publication matters. A mention in an industry journal matters. A mention in a podcast transcript, if it's published on a reputable site, matters. What doesn't matter much is a customer review on your own site or a testimonial you write yourself. The model learns to discount self-published material when evaluating credibility.
If you run a boutique professional services firm and you want ChatGPT to recommend you, you need to be featured in publications that OpenAI's training data includes. That might mean contributing articles to industry outlets, getting quoted in trade journals, or appearing in case studies published by recognised sources. It's PR and content strategy, not technical SEO.
The timeline also matters. If most of the trustworthy mentions of your business are more than two or three years old, ChatGPT's model may deprioritise you. Active, ongoing mentions in reputable sources signal that you're still real and still relevant. A company that was written about five years ago and hasn't been mentioned since looks like it might no longer be operating at the level the model learned about.
Authority Signals and Expert Networks
ChatGPT weights recommendations higher when they come from networks of authority. If you're mentioned in a prestigious business publication, and that publication is cited and referenced across other authoritative sources, the model treats that mention as more significant. It's not enough to be mentioned; the sources that mention you need to have their own authority signals.
This is different from Google, which looks at link equity. The AI model is looking at textual patterns and context. It's asking: "Where does this entity appear? In what kinds of sentences? Alongside what other entities? In what types of publications?" If your business is mentioned in articles that also mention industry peers, regulatory bodies, and market leaders, you appear more authoritative.
One practical implication: partnerships and affiliations matter more than before. If you're affiliated with a well-known industry body, or if you've partnered with an established player, and that partnership is documented in published sources, your entity gains authority by association. You don't need to be famous; you need to be credibly connected to things the model already knows are authoritative.
The Role of Reviews and Ratings
While ChatGPT doesn't have access to live review platforms like Google Reviews or TripAdvisor, the reviews and ratings on those platforms do influence training data in subtle ways. Major publications often report on businesses with strong ratings. High-rated businesses get mentioned in recommendation articles. Those articles then become part of training data, and suddenly your high ratings translate into being recommended by AI.
The indirect path matters more than direct access. A business with 4.8 stars and 500 reviews will be mentioned more often in published articles and recommendation posts. Those mentions, in context, signal to the model that this business is well-regarded. You want reviews not just for Google or for customers, but because they increase the likelihood that publications will write about you, and those publications become training data for AI models.
This also means that recent, authentic reviews carry more weight than older ones. If you're getting consistent 4+ star ratings in the recent months, you're more likely to show up in ChatGPT recommendations because more recent articles will mention you. If your reviews are all from three years ago, you look stagnant.
What You Can Control
You can't control ChatGPT's training data directly. You can't pay OpenAI to rank you. But you can make your business more visible and authoritative in the sources that ChatGPT learned from. That means a real content strategy—getting features in publications your customers read, contributing expertise to industry conversations, building partnerships with complementary businesses, and maintaining consistent, professional listings across platforms.
It also means being very intentional about your entity definition. Know what you call yourself. Know what vertical you're in. Know who your peers are. Make sure every public appearance of your business is coherent, consistent, and credible. The AI isn't reading your site with human judgment. It's looking for patterns. Give it clear patterns to learn.
Most businesses are still optimising for Google, publishing content to Google Search, and hoping for backlinks. By the time they look up from that work and notice that ChatGPT isn't recommending them, they've lost eighteen months of visibility. Start now. Get into publications. Build real authority. Make your business something the model has learned to trust. That's how you stay recommended in the AI era.
— Sam