ChatGPT SEO: how to show up when customers ask ChatGPT

How does ChatGPT decide which businesses to recommend?

ChatGPT builds recommendations from two sources: its training data, which reflects what the open web has said about businesses over years, and live web search, which it runs mid-conversation for current or local questions. When both paths agree on who you are, what you do and where you are, the model names you with more confidence. When they disagree, it hedges or skips you.

One caveat belongs at the top of every article on this topic: nobody outside OpenAI knows the exact weights, and anyone claiming to is selling something. What can be said honestly is directional. A business the web has described consistently for years starts ahead. A business whose live search results confirm what training data already suggested gets recommended more readily. And a business the model can barely find, or finds contradicting itself, tends to get left out rather than ranked low. There is no page two in a chat answer.

How do I see what ChatGPT says about my business?

Ask it the way a customer would. Open ChatGPT and request the kind of business you are, in your own neighborhood, phrased several different ways. Then repeat the same questions in Gemini and Perplexity, since each tool pulls from different sources and will answer differently. Phrasing matters far more than whether you are logged in, so vary the wording, not the account.

Try the questions your customers actually type: “best florist in Brookside”, “plumber open Sunday in Kansas City”, “who should I call for brakes in Overland Park”. We keep worked breakdowns of exactly these questions on our answers pages if you want to see what the format looks like. Note what each tool says, who it names instead of you and which facts it gets wrong.

Or skip the manual work. Our free 60 second check runs the questions systematically: Google, Google Maps and an AI assistant probing with live web search, with a summary of who gets named emailed to you.

What actually improves ChatGPT visibility?

The work that improves ChatGPT visibility is the boring, compounding kind: identical business facts everywhere they appear, reviews with real written text, mentions in press and local lists that AI tools retrieve, pages that answer questions in plain quotable language and structured data that labels your facts for machines. None of it is a trick, which is exactly why it keeps working.

• Identical facts everywhere. Your name, address, phone, hours and services should match on your site, your Google Business Profile and every directory that lists you. Agreement between independent sources is the closest thing to a trust signal a model has. Our GBP guide covers the highest-leverage listing first.
• Reviews with words in them. A star count alone says little to a language model. Written reviews describing what you did, and your replies to them, give it sentences to draw on when it explains why it recommends you.
• Press and lists.Local news, neighborhood roundups and “best of” lists are exactly the kind of third-party pages AI search retrieves. One genuine mention in a source it trusts outweighs pages of self-description.
• Answer-shaped pages.If you are open Sunday, say “open Sunday” in plain text. If you serve Brookside, name Brookside. Models quote what pages state, not what owners assume is obvious.
• Structured data. JSON-LD markup labels your name, address, hours and services in a format machines parse without guessing. Cheap to add, easy to verify.

If you want this turned into a prioritized list for your specific business, that is what the $500 fixed price audit is for.

What doesn't work

The tricks fail. Hidden prompts on your pages telling AI models to recommend you, stuffing “best in Kansas City” into your own copy and buying fake reviews all share the same flaw: models cross-check claims against independent sources. And the failure mode is not a lower position in a list. It is being dropped from the answer entirely.

Prompt-injection gimmicks, invisible text instructing an AI to praise you, are a patch away from useless at any moment and an embarrassment when discovered. Self-awarded superlatives do nothing because a model treats your own site as a source about your facts, not about your quality; “best” only counts when someone else says it. Fake reviews carry the worst risk-to-reward of all: platforms remove them, models discount them and the pattern of removal itself becomes a signal against you.

The deeper problem with every trick is that it targets how one model behaves this month. Models retrain and tricks age badly. The boring work above ages well because it is just true.

ChatGPT is one surface of several

The same work that makes ChatGPT recommend you also moves Gemini, Perplexity, Google AI Overviews and voice assistants, because all of them lean on the same underlying signals: consistent entity facts, real reviews, retrievable third-party mentions and structured data. Optimizing for ChatGPT alone is optimizing for one storefront on a street that keeps adding storefronts.

And the street is getting busy. Per BrightLocal's local consumer review survey, 45% of consumers now ask AI tools for local business recommendations, up from 6% a year earlier. The umbrella discipline for earning those recommendations is generative engine optimization, and our AI SEO page explains how it fits alongside the classic search work you may already be doing.