If you’ve typed your own business name into ChatGPT and gotten a confident, accurate answer, then asked the question a stranger would actually type — best [your service] near me — and gotten someone else’s name back, you’ve already found the pattern this page exists to explain. It shows up often enough in our own audits that it’s worth explaining properly, not glossing over with a sales pitch.
Why doesn’t ChatGPT recommend my business, even though it clearly exists?
Because the engine isn’t running a fresh search when it answers — it’s drawing on what it has already read and decided to trust, and if your business has never appeared anywhere in that material in a form worth quoting, there’s nothing for it to cite. This is the single most common misunderstanding we see: assuming an AI engine works the way a search engine did, crawling and re-ranking a live index at the moment someone asks. It doesn’t. Ask it something and it’s synthesizing an answer from a citation graph — your own site, directories, reviews, forum threads, trade coverage — that it already trusts, weighted toward sources it has seen cited elsewhere. A business can be real, licensed, reviewed, and fifteen years in operation and still be invisible in that graph, for the same reason a book nobody has ever quoted doesn’t show up in anyone’s bibliography — not because it’s wrong, because nobody’s referenced it yet.
What’s the pattern behind “AI knows my name but recommends someone else”?
Across every audit in our own book, the split is the same: 75–100% presence when an engine is asked about the business by name, and 0–10% presence when it’s asked the question a real buyer would actually type — best [service] near me, who fixes [problem], same-day [service] in [city]. Six audits, six times, the same shape. That’s not a coincidence and it’s not six unlucky businesses — it’s what happens when a real, reputable business has real name recognition but nothing on the discovery side of the citation graph: no page that answers the buyer’s actual question, no presence in the third-party sources an engine leans on when the buyer hasn’t already decided who to call. Brand recall and discovery visibility are two different things an engine has to earn separately, and most businesses have only ever built the first one.
What do you actually look at first, and why that order?
We start with the gap the split above describes: what would a stranger, not someone who already knows your name, need to read before an engine could confidently name you. That usually means checking whether your site answers the specific questions buyers ask — not just describes your services in general terms — and whether the third-party sources engines already trust for your category have you listed at all, correctly. Technical fixes come after that, because a perfectly structured page with nothing to cite doesn’t move the needle, and a messier page that already answers the right question sometimes does. It’s the same order we used on ourselves before we used it on anyone else — we published that result before we ever ran the method on a client — and it’s what a full audit turns into a sequenced, prioritized list for your specific business.