Guide
How to check if ChatGPT recommends your business (and what to do if it names a competitor)
Most owners have tried this once: opened ChatGPT, typed their company name, seen a passable answer, and moved on. That test tells you almost nothing. Buyers who have never heard of you can’t ask about you by name — the question that matters is the one they actually type, with no brand in it. This guide shows you how to run that test yourself, by hand, for free, and how to read what comes back. It takes about an hour.
The manual method
Step 1Write the question your buyer would ask
Buyers who haven’t heard of you can’t ask about you by name. Asking an engine about your brand tests whether it knows you exist; asking the buyer’s question tests whether it would recommend you. Write the question the way a real buyer would type it, with the problem, the constraint, and the place: “Who are the best employment lawyers in Chicago for a 50-person company?” If location matters to your buyers, put the city or region in. Write the wording down exactly — you’ll reuse it verbatim, because changing the wording can change the answer.
Step 2Run it in ChatGPT, Gemini, and Perplexity
Paste the identical question into all three engines, logged out or in a private window so your own history doesn’t shape what comes back. The three engines answer differently and show different sources, which is exactly why checking one engine isn’t enough. Save each full answer — copy the text into a document rather than screenshotting the top of the page.
Step 3Record three things from each answer
First: is your business named at all? Second: who is named instead — which competitors, and in what position? Third: which sources are visible — the links, directories, review sites, or articles shown alongside the answer. The sources are your working clue about where the answer came from.
Step 4Repeat on a different day
AI answers change with the model, the wording, your location, and the date, so one run is a snapshot, not a ranking. Run the same question again a few days later. Being named once in three tries is a weak, unstable position. A competitor named every time is a stable one — and that’s the gap that matters.
Step 5Read the sources like a buyer would
Open every visible source. Could your business plausibly appear there? If the engines keep citing a directory you’re not listed in, an association you’re not a member of, or a review site with no reviews of you, you’ve found a concrete gap — the kind you can actually fix.
Step 6Turn the gaps into a short fix list
Most fixes fall into four buckets: your own site (clearer service pages, direct answers to buyer questions), third-party profiles (directories, reviews, associations), community presence (the Reddit threads and forums engines quote), and digital PR (being cited in articles engines lean on). Rank the list by effort and likely impact, then start at the top.
An honest note on app vs. API results
Results in the ChatGPT or Gemini apps can differ from what the underlying APIs return. The apps personalize, carry your chat history, and may search the web in ways a clean API run doesn’t replicate. Neither is “the truth” — both are samples. CiteMe runs the API versions of all three engines with live web search; the exact models and settings are documented on the methodology page.
If a competitor gets named
Don’t panic, and don’t rewrite your homepage overnight. First check the obvious: does the competitor show up in places you don’t? Look at their service pages — do they answer the buyer’s question directly, in plain language, near the top of the page? Then look at the sources the engines showed. If the same directory or review site keeps appearing, getting properly listed there is usually the highest-leverage first move.
Next, look at your own site the way an engine does. Could a crawler tell, from your pages alone, what you do, who it’s for, and where? Vague taglines and brochure copy give an engine nothing to repeat. A short, direct answer block — the buyer’s question, answered in two or three sentences — gives it something to quote. So does an FAQ page written in the real wording of buyer questions.
Finally, accept that some answers move slowly. Engines lean on sources built up over years, and a single new page rarely flips a recommendation on its own. The goal isn’t to trick a model — it’s to be accurately described in the places the model already reads, so that when a buyer asks, there’s something solid to find.
Patterns, not screenshots
One more discipline: track patterns, not moments. A single answer on a single day proves very little. What matters is whether you’re named consistently, whether the same competitors keep showing up, and whether the sources shift as you fix things. Write down the question, the date, the engine, and the result each time, and a month of casual checks becomes real evidence. It’s also why every CiteMe result shows the exact question, engines, date, and sample size — a result you can’t re-run is a result you can’t trust.
Or skip the manual work: run one question through all three engines at once with the free snapshot. You write the question, CiteMe runs it through ChatGPT, Gemini, and Perplexity, and you get each engine’s full answer and visible sources as a PDF.