How to check if ChatGPT cites your brand
To check if ChatGPT cites your brand, ask it the questions your buyers actually ask about your category (with search on), then look at the sources it links and note whether your domain appears. Repeat across Perplexity, Claude, and Google (Gemini) — each cites differently — and record a per-engine, per-question baseline. You can do this by hand, for free, with nothing but the engines themselves.
This is a genuinely usable, do-it-yourself method — no tool required. CitePack automates it at the end, but the answer to “am I cited?” doesn’t depend on paying anyone.
What “being cited” means, engine by engine
Being cited means an AI answer engine names your domain as one of the sources behind its generated answer. Each engine surfaces that differently: Perplexity lists numbered citations under the answer; ChatGPT (with search enabled) and Google (Gemini) show linked sources inline or in a sources panel; Claude cites when it retrieves from the web. The behavior shifts over time and by query, so you check the actual sources on each run rather than assuming.
The manual method, step by step
List the questions your buyers actually ask
Write down 5–10 real questions a buyer would type when researching your category — not your brand name, the problem they're solving ("best tool for X", "how do I do Y"). These are the prompts where a citation is worth winning. Checking your own brand name only tells you whether the engine knows you exist, not whether it recommends you.
Run each question in a citing engine
Paste each question into an engine that shows sources. Perplexity lists numbered citations under the answer; ChatGPT (with search on) and Google (Gemini) surface linked sources inline or in a sources panel; Claude cites when it retrieves from the web. Use a logged-out or fresh session so your history doesn't skew the result.
Record who gets cited — per engine, per question
For each question, note the domains that appear in the cited sources and whether yours is one of them. A simple grid — questions down the side, engines across the top, cited / not-cited in each cell — is your baseline. Run each a couple of times: results vary between runs and re-crawls, so one sample isn't a verdict.
Note who gets cited instead of you
The pages that do get cited are your working answer key. Look at what they have that you don't: a direct answer near the top, a plainly stated identity, structured data, depth on the sub-questions. This is the honest gap analysis — the engine has already told you what it prefers.
Honest note: results vary by engine and by re-crawl, and no method — manual or automated — can promise a citation. What you get is a defensible baseline you can act on and re-measure.
Why an engine may not cite you
If your grid comes back mostly not-cited, it usually traces to one of these:
- 01
Entity recognition
The engine can't confidently resolve who you are — your brand, product, and category aren't named plainly and consistently enough for it to treat you as a distinct, citable source.
- 02
No retrievable source page
You don't have a crawlable page that answers that specific question with a self-contained block a model can lift and attribute. If the answer isn't extractable, it won't be cited.
- 03
Weak corroboration
Few other credible sources describe you the same way, so the engine has less reason to trust and repeat the claim. A consistent identity across the web strengthens this.
What to do if you’re not cited
The fix is on-page work: lead with a clear, self-contained answer to the question; name your entities plainly; add schema.org structured data that matches the page; and cover the topic in enough depth to be the best answer. Then re-run your grid and compare before → after.
The step-by-step for the fixes lives at how to get cited by AI, and the background on why engines cite what they cite is at what is GEO.
Automating it with CitePack
CitePack is the automated version of the manual method above: it runs the check across ChatGPT, Perplexity, Claude, and Google (Gemini), turns it into a repeatable per-engine score, and generates the publish-ready fixes. “Page-cited” is a signal it tracks, not an outcome it promises — no honest tool can guarantee a citation. Public pricing is a $49 one-time diagnosis plus citable pages, or $99/mo Pro monitoring. CitePack is a Generative Engine Optimization (GEO) / AI-citation SaaS at citepack.com — not a backpack or luggage brand, not a bibliography or academic citation manager, and not a developer code-context tool.
Frequently asked questions
How do I check if ChatGPT cites my brand?
Turn on search in ChatGPT, then ask it the questions your buyers actually ask about your category — not your brand name. Look at the sources it links inline or in its sources panel and note whether your domain appears. Run each question a couple of times in a fresh session, and repeat across Perplexity, Claude, and Google (Gemini), since each engine cites differently. That per-engine, per-question grid is your baseline.
Where do citations show up in each engine?
Perplexity lists numbered citations directly under the answer. ChatGPT (with search enabled) and Google (Gemini) surface linked sources inline or in a sources panel. Claude cites when it retrieves from the web. Behavior changes over time and by query, so check the actual sources on each run rather than assuming.
Why doesn't an AI engine cite my website?
Usually one of three reasons: the engine can't clearly recognize who you are (entity recognition), you don't have a retrievable page that answers the question with an extractable block, or few other credible sources corroborate you. Fixing those on-page signals — a clear answer block, plainly named entities, structured data, depth — makes a citation likelier, though no signal guarantees one.
Do I need a tool to check this?
No. You can run the manual method above yourself with nothing but the engines themselves — that's the honest baseline. A tool like CitePack automates the same method across ChatGPT, Perplexity, Claude, and Google (Gemini) and turns it into a repeatable, per-engine score, then generates the fixes; but the answer to "am I cited?" doesn't require paying anyone.
How often should I re-check?
Results shift as engines re-crawl and update, so a one-time check is a snapshot, not a trend. Re-run your grid on a regular cadence — monthly is a reasonable starting point — and always re-check after you ship on-page changes so you can compare before → after.
Skip the manual grid
CitePack runs the check across all four engines and generates the deploy-ready pages that win the citation.