Used vs Cited
The two ways a brand appears in AI search. 'Used' means your information shapes the answer with no attribution; 'Cited' means you are named as a clickable source. Tracking only citations misses half of your visibility.
What is Used vs Cited?
Used vs Cited distinguishes the two ways a brand shows up in an AI answer. When your brand information helps produce the answer but carries no source label, that is "Used." When your brand is named as a clickable link or source card, that is "Cited."
Being used is like Google indexing a page before it ever ranks it: the AI engine consumes your information to build the answer but returns no credit to the reader. Being cited is the visible form — a source card or in-body link the user can see and click.
| Mode | Definition | Visible to the user? |
|---|---|---|
| Used | Your information shapes the answer body, but with no source label | ✕ Unattributed — no credit, no click |
| Cited | Your brand is named as a clickable source or link | ○ Attributed — click-through possible |
Why citations alone are not enough
Citations are only part of AI visibility. Because AI systems weave brand information into answers without naming sources, tracking citations alone leaves roughly half of your real influence unmeasured.
Per Ahrefs analysis, ChatGPT pulls almost the same number of cited (16.57) and uncited (16.58) URLs per response. In other words, about half of the information the model consumes is absorbed into the answer without attribution.
Unattributed information also skews toward certain platforms. In the same 1.4M-prompt analysis, Reddit accounted for 67.8% of uncited sources — evidence of a platform bias in which content gets consumed without credit.
There are traffic implications too. Pew Research found an 8% click-through rate on searches with an AI summary present, versus 15% without — roughly half. Yet the quality of that traffic differs: Similarweb data reported an 11.4% conversion rate for AI-referred visits against 5.3% for organic search. Fewer clicks convert to visits, but those visitors arrive with higher intent.
Measuring both axes
Used and cited are captured by different metrics.
- Used — whether your brand name appears as text in the answer body. → Brand Mention Share
- Cited — your share of the source cards and links. → Citation Share
Reading both axes together separates the "used in the body but missing from sources" case (Used > Cited) from the "listed as a source but barely shaping the answer" case (Cited > Used). The two call for different fixes.
| Observation | Diagnosis | Prescription |
|---|---|---|
| High Used · Low Cited | Authority exists but is not credited as a source | Strengthen robots.txt, structured data, extraction structure |
| Low Used · Low Cited | No entry at all | Build earned media and category entry points |
| High Cited · Low click | Exposed but not converting to visits | Sharpen brand imprint and differentiation inside the answer |
The Citation Selection vs Absorption framework — separating the stage where a source is picked from the stage where it is absorbed into the answer body — sharpens this diagnosis further. And usage consumed without a source tends to become Dark Traffic that never shows up in your referral logs.
Related terms
Is your site visible in AI search?
See for free how ChatGPT, Perplexity, and Gemini describe your brand.
Start Free Diagnosis →