How to Track Your Brand's Citations & Mentions in ChatGPT, Claude & Perplexity (2026)

To find out whether AI engines cite your brand, run the questions your buyers actually ask into ChatGPT, Perplexity, Claude, and Google AI Overviews, then log two things in each answer: whether your brand is cited (named with a linked source) or merely mentioned (named with no attribution), and which competitors and source URLs appear instead of or alongside you. The catch is that AI answers are stochastic — the same query returns different sources on different runs — so a single check tells you almost nothing. Reliable tracking means sampling the same query set repeatedly and watching "share of answers" as a trend, which is what automated tools (including LumenGEO's free audit) exist to do for you.

Most people start here after a hunch: a prospect says "I asked ChatGPT and it recommended your competitor," or referral traffic from perplexity.ai shows up in analytics. You want to know if it's real, how often it happens, and what to do about it. This guide gives you the exact manual method for each major engine, explains why manual checking alone is unreliable, and shows what to actually track.

Last updated: June 2026

What's the difference between a brand citation and a brand mention in AI answers?

A citation is an explicit, linked source — the AI names your brand and attributes a claim to your page, usually with a clickable footnote or source card. A mention is your brand named in the prose with no source attached. Both shape buyer perception, and both are worth tracking, but they mean different things.

A citation says the engine retrieved your content, judged it credible enough to ground a claim on, and surfaced the link. That's the highest-value outcome: it can drive a click, and it signals retrieval authority. A mention — "popular options include Acme, Brand X, and Brand Y" — means your brand is part of the model's learned or retrieved understanding of the category, even without a source. A mention with no link still influences a shortlist; it just won't show up as referral traffic.

The practical implication: don't only chase the linked footnote. If competitors are mentioned in the answer body and you aren't, you have a presence problem even before you have a citation problem. Track both columns.

How do I check if ChatGPT cites my brand?

Use ChatGPT with search/browsing active, run your real buyer queries, and read the sources list under the answer. ChatGPT retrieves through Bing as its backend, so the sources it surfaces are pulled from Bing's index — which is also why Bing Webmaster Tools is a useful free signal for ChatGPT visibility.

Concrete steps:

1. Start a fresh chat and make sure web search is on (look for the search/globe toggle, or phrase the query so it triggers retrieval — "what are the best ... in 2026" reliably does).
2. Ask the question the way a real buyer would, not the way a marketer would. "Best [category] tools for [use case]" beats "[your brand] review."
3. Read the answer body for mentions (is your brand named in the prose?) and expand the sources list for citations (is your domain among the linked sources, and at what position?).
4. Log the result: cited / mentioned / absent, your position, and every competitor and source URL that appeared.
5. Re-run the same query several times — you'll see the source set shift between runs. That shift is the point of the next section.

Note that source URLs are often not your competitors' homepages. They're frequently Reddit threads, listicles, G2, and review sites. If a Reddit thread is cited and your brand isn't in it, that's an actionable gap.

How do I check if Perplexity cites my brand?

Perplexity is the easiest engine to audit because it is citation-dense and shows sources prominently — numbered inline footnotes plus a sources bar at the top of every answer. Run your buyer queries, then read the numbered citations directly; there's no hidden retrieval to coax out.

Perplexity cites heavily: in LumenGEO's analysis, a typical Perplexity answer surfaces roughly 21.87 sources. That density means more slots to win — and more competitors visible per answer. It also means the source mix is revealing. Around 46.7% of the top-10 sources Perplexity pulls are Reddit, so if your category lives on Reddit and you're absent there, Perplexity will struggle to cite you regardless of how strong your own site is.

Steps are the same shape as ChatGPT: ask the buyer query, read the inline footnotes for citations, scan the prose for unattributed mentions, log your position, and capture every source URL. Then re-run. For a deeper Perplexity-specific workflow, see Perplexity SEO: how to get cited.

How do I check if Claude cites my brand?

Use Claude with web search enabled, ask your buyer query, and check the linked sources Claude returns. Claude only cites external sources when it actually performs a web search, so make sure the query triggers retrieval (recency- or recommendation-shaped questions are most reliable).

Read the answer for both mentions in the prose and the linked sources Claude surfaces, log your position and the competitors/URLs that appear, and re-run. Treat Claude as its own trend line — don't assume that being cited in ChatGPT means you're cited in Claude. They draw on different retrieval paths and surface different sources.

How do I check if Google AI Overviews and Gemini mention my brand?

For Google AI Overviews, run your buyer query in Google Search and look at the AI-generated answer block that appears above the classic results, including its linked source cards. For Gemini, ask the same query in the Gemini app with its sources expanded. Log the same columns: cited, mentioned, position, competitors, source URLs.

A measurement caveat worth knowing: clicks from Google's AI surfaces don't get their own bucket — they report into Search Console's general "Web" data, so you can't cleanly separate AI-Overview-driven clicks from regular organic clicks. That's one reason citation presence (what the answer says) is a more honest metric here than click attribution. More on that in how to track AI search traffic.

Why is manually checking AI citations unreliable?

Because AI answers are stochastic: ask the identical question twice, a few minutes apart, and you'll usually get a different set of cited sources. A single manual check is one random draw from a distribution — seeing "not cited" once tells you almost nothing, the way one coin flip tells you almost nothing about the coin.

Three things drive the variation. Retrieval samples from a candidate pool rather than returning a fixed list. The query often fans out into sub-queries that differ between runs. And generation itself is probabilistic. Stack those together and the source list legitimately changes run to run, even with no change to your site or the web.

This is why a one-off "I checked and we're not in ChatGPT" is not a finding — it's anecdote. To turn it into a measurement, you sample.

Repeated sampling and "share of answers"

The reliable approach estimates the distribution instead of sampling it once:

• Use a fixed query set. Define 20–30 queries that matter to your business — category, comparison, and problem queries phrased the way a real person asks an AI — and measure that same set every period. A changing query set makes period-to-period comparison meaningless.
• Run each query many times. 10–20 runs per query is a reasonable working standard. Cited in 6 of 20 runs is roughly a 30% citation rate for that query — a real number, where a single check is not.
• Report share of answers. Across your full prompt set and all runs, how often does your brand appear? It's a percentage over many observations, so it moves slowly and meaningfully. Share of answers moving from 18% to 24% over two months is a result; "ChatGPT cited us today" is not.
• Measure per engine. Only about 11% of the domains cited by ChatGPT are also cited by Perplexity, so a blended number hides more than it reveals. Run each engine as a separate trend line.

What should I actually track for brand citations in AI?

Track four things, per engine, over time: citation presence, prominence, competitor citations, and the source URLs the engine pulled. Logging only "yes/no cited" throws away most of the signal.

The source-URL column is the most underused and the most actionable. If Perplexity keeps citing a specific Reddit thread or a "best tools" listicle that omits you, you now know exactly where to earn a presence — getting included in that listicle or genuinely participating in that thread can do more for your AI citations than another blog post on your own domain.

Pair this with a single overall metric so you can see the forest, not just the trees — see what a GEO score is for how presence, prominence, and competitor data roll up into one number.

What tools track brand citations across AI engines automatically?

Automated monitoring tools exist precisely because manual repeated sampling across multiple engines is tedious — they run your query set, sample over time, and report share of answers and competitor data for you. They fall into two camps: a free starting point, and dedicated monitoring dashboards.

LumenGEO is the option built around fixing the gap, not just watching it. The free audit checks your brand's citations across ChatGPT and Perplexity, shows which competitors are getting cited instead of you, and surfaces the source URLs behind the answers — so you get the manual method's findings in one pass instead of dozens of runs. Paid monitoring then samples that query set over time so you can read share of answers as a trend rather than a snapshot, and the GEO action plan turns the gaps into prioritized fixes.

If you want a pure monitoring dashboard, other tools focus there — Profound, Otterly, and Peec are the names you'll encounter. Rather than re-review each one here, we keep a maintained comparison of Perplexity citation-tracking tools so you can match a tool to your budget and engine coverage.

The honest framing: a tool isn't required to find out whether you're cited — the manual method above does that for free. A tool is what makes the measurement repeatable and trended without you running hundreds of queries by hand, and what turns "we're not cited" into "here's who is, and here's the gap."

A short closing note

Checking whether AI engines cite your brand is genuinely worth doing — but do it honestly. One run that omits you isn't a verdict, and one run that includes you isn't a win. Run your real buyer queries across each engine, log citations, mentions, competitors, and source URLs, and read the result as a trend over repeated samples. That's the difference between an anecdote and a measurement.

If you'd rather see your current citation picture across ChatGPT and Perplexity — including which competitors are cited instead of you — in a couple of minutes instead of an afternoon of manual runs, run a free audit.

Frequently Asked Questions

What's the difference between a citation and a mention in an AI answer?

A citation is your brand named with an explicit, linked source — the AI attributes a claim to your page, usually via a clickable footnote or source card. A mention is your brand named in the answer's prose with no source attached. Citations can drive clicks and signal retrieval authority; mentions still influence a buyer's shortlist even without a link. Track both, per engine.

How do I check if ChatGPT cites my brand?

Start a fresh ChatGPT chat with web search on, ask the question a real buyer would ask ("best [category] for [use case]"), then read the answer body for mentions and expand the sources list for linked citations. Note your position and every competitor and source URL that appears, then re-run the query several times — the source set will shift, which is why one check isn't enough.

Why do I get different AI sources every time I ask the same question?

Because AI answers are stochastic. Retrieval samples from a candidate pool, the query often fans out into varying sub-queries, and text generation is probabilistic. Together these make the cited sources legitimately change between runs even with no change to your site. That's why a single check is one random draw, and why reliable measurement uses repeated sampling.

How many times should I run a query to measure AI citations?

A reasonable working standard is 10–20 runs per query across a fixed set of 20–30 buyer queries, repeated each measurement period. Cited in 6 of 20 runs is roughly a 30% citation rate for that query — a real measurement. Report the result as "share of answers" per engine and read it as a trend over rolling windows.

Do I need to track each AI engine separately?

Yes. Only about 11% of the domains cited by ChatGPT are also cited by Perplexity, so engines surface meaningfully different sources for the same query. A blended number hides which engine you're winning or losing. Run ChatGPT, Perplexity, Claude, and Gemini/Google AI Overviews as separate trend lines.

What's the best free way to check brand citations in AI?

The manual method costs nothing: run your buyer queries across ChatGPT, Perplexity, Claude, and Google AI Overviews and log citations, mentions, competitors, and source URLs. For a faster pass, LumenGEO's free audit checks your citations across ChatGPT and Perplexity and shows which competitors are cited instead of you. Bing Webmaster Tools is also useful for ChatGPT visibility, since ChatGPT retrieves through Bing's index.

Why do Reddit and review sites keep showing up as my competitors' sources?

Because AI engines cite the pages they judge most credible for a query, which are often third-party — Reddit, G2, listicles, and review sites — rather than brand homepages. Around 46.7% of Perplexity's top-10 sources are Reddit. If those pages cite competitors and not you, earning a genuine presence there can do more for your AI citations than another post on your own domain.