AI engines name your brand in three different ways, and most marketers only count one of them. A citation is when the AI names you and links you. A mention is when the AI names you without a link. A link is when the AI uses you as a source without naming you. All three matter for revenue, but they generate different signals and demand different optimization moves. Conflating them is one of the most common — and most expensive — mistakes in early GEO programs.
This page is the clean definition of each, the data on how often each occurs, and how to track all three.
The three outcomes, defined
| Outcome | What the user sees | Does the brand get named? | Does the brand get linked? | Typical share of brand appearances |
|---|---|---|---|---|
| Citation | "Brand X makes a strong vitamin C serum…" | Yes | Yes | ~15% |
| Mention | "Brand X makes a strong vitamin C serum…" | Yes | No | ~70% |
| Link | "Vitamin C serums in this price range tend to use 15% L-ascorbic acid<sup>[1]</sup>…" (where [1] links to your site) | No | Yes | ~15% |
Approximate shares vary by engine, query type, and category — but across multiple 2025–2026 studies, the broad pattern holds: mentions dominate, citations are rarer than people think, and source-only links are roughly as common as citations.
One study found that 85% of brands named in ChatGPT answers have no accompanying citation link. That's the most important number on this page. Most AI brand exposure is namedrop, not clickable link.
Why this matters for measurement
If you're measuring AI visibility by "did my domain appear in a citation," you're missing roughly 85% of the value AI search is delivering. A brand named twenty times across ChatGPT responses without ever being linked is doing far more for downstream revenue than a brand cited once but never recommended by name.
The right way to measure AI visibility is three metrics, tracked separately:
- Citation rate — % of relevant AI responses that include a clickable URL to your domain
- Mention rate — % of relevant AI responses that name your brand (with or without a link)
- Source rate — % of relevant AI responses that use your domain as a source (with or without naming you)
You want mention rate as your primary metric for brand visibility. Citation rate as your secondary metric for traffic and authority. Source rate as a leading indicator of how much AI engines trust your content.
What each outcome tells you
Citation (named + linked)
The most valuable single outcome. The AI both recommends your brand by name and gives the user a clickable path to your site. Citations are rare because they require both:
- The AI engine considers your brand directly relevant to the query (which earns you the name)
- The AI engine considers your domain authoritative enough to cite as a source (which earns you the link)
Citations drive direct traffic and the strongest brand signal. They also tend to convert well — a user who clicks an AI citation has typically already been pre-sold by the AI's framing of your brand.
How to earn more citations: First-party content depth. Schema.org markup. Authority on your own domain. Brands cited frequently are usually the brands publishing the most fact-dense, well-structured content in their category. Citation rate is the slowest metric to move, but it's the most durable.
Mention (named, not linked)
The most common outcome and the most underrated. The AI recommends your brand in the answer but doesn't give a clickable path. The user might Google your brand next, or remember it for next time, or screenshot the answer — but they don't click through to you directly.
Mentions matter because they shape consideration sets. A brand mentioned three times across an AI shopper's research session enters their consideration set even if they never click. That's exactly the kind of upstream brand-building marketers used to pay for with TV and influencer content — and AI search is now delivering it for free, to the brands that earn it.
How to earn more mentions: Third-party corroboration. Reviews, publishers, communities, and forums. AI engines name brands they've seen consistently across multiple sources — even if those mentions don't link to the brand directly. This is also why brands with strong word-of-mouth in their category tend to dominate AI mentions even when their site SEO is weaker than competitors.
Link (used as source, not named)
The least visible outcome and the most misunderstood. The AI uses your content as a source — adding it to the answer's sources panel or footnoting it — but doesn't name your brand in the answer itself.
This happens most often when:
- Your content is treated as a reference (a definition, a stat, a study) rather than a recommendation
- Your content is generic to the category rather than specific to your brand
- The AI uses your data without endorsing your brand
Links generate referral traffic but limited brand impact. They're a strong signal that AI engines trust your content — which is leading indicator of citations to come — but they don't move purchase consideration directly.
How to earn more source-only links: Publish original data, frameworks, and definitions that other content in your category needs to reference. Research reports, benchmarks, taxonomies, ingredient databases, calculators. The kind of content that gets cited as a source rather than recommended as a brand.
The three-by-five matrix
Each of the five major AI engines behaves slightly differently across these three outcomes. The pattern roughly looks like this:
| Engine | Citation rate (relative) | Mention rate (relative) | Source-only link rate (relative) |
|---|---|---|---|
| ChatGPT | Lower | Highest | Medium |
| Claude | Medium | Medium | Medium |
| Perplexity | Highest | Medium | Highest |
| Copilot | Medium | Medium | Medium |
| Gemini | Higher | Lower | Medium |
What this tells you in practice:
- If you want named recommendations, prioritize ChatGPT. It mentions more brands without linking than any other engine, but those mentions move consideration.
- If you want clickable traffic, prioritize Perplexity. It cites more aggressively and links more aggressively than the others.
- If you want to be used as a trusted source, publish original data. Perplexity uses sources most heavily, but all five reward genuinely original content.
How to track all three at scale
Most marketing analytics tools weren't built for this. They count clicks, not mentions. Even the AI search monitoring tools that exist today often focus on citations only, missing the mention surface.
The right setup tracks three things per query:
- Did the AI engine name your brand in the response? (mention)
- Did the AI engine include your domain as a clickable citation? (citation)
- Did the AI engine include your domain as a source/reference without naming you? (source-only link)
Run that measurement across a defined set of 25–75 priority prompts per category, refreshed weekly. Track the trend over time per engine. Cross-reference against actual Shopify orders to get attribution-grade revenue per AI-engine event.
RevvUp.ai does this natively for Shopify brands across all five major engines. But the framework matters more than the tool — if you set up the measurement properly with any tool (or manually, if you have to), you'll catch the mention surface that most teams miss.
What this changes about your strategy
Three concrete shifts most brands need to make:
- Stop reporting "AI citations" as your primary metric. It's understating your true visibility by 5–7x. Report mention rate first, citation rate second.
- Invest more in third-party corroboration than first-party content. Mentions are earned mostly off your own site. Reviews, publisher coverage, community presence, and expert endorsement drive mention rate more than on-site optimization does.
- Track separately by engine. A unified "AI visibility score" hides the fact that ChatGPT and Perplexity behave very differently. Per-engine dashboards force the team to optimize for the engine that matters most in your category.