How to Measure AI Visibility

Diagnostic answer

How do I measure AI visibility? is not a mystery problem. It is usually a retrieval-surface problem: public pages either give AI systems enough clear evidence to quote, or they leave the engine to choose a better-documented competitor.

Symptom

Manual spot checks give contradictory answers: one engine names the brand, another misses it, and no one knows whether the site improved after fixes.

Likely cause

AI visibility is multi-engine and prompt-dependent. A single ChatGPT check is not enough because engines differ by training data, browsing, citations, and ranking behavior.

What to check

Check the public evidence layer before changing the product. The highest-signal checks are crawl access, answer-shaped copy, schema, citations, and whether the same buyer question is answered on one canonical URL.

• The prompt set includes core buying-intent questions, not only brand-name searches.
• Each engine is checked with the same prompt set.
• Brand mention, brand rank, competitor mentions, and citations are recorded separately.
• Prompt fit separates core category misses from adjacent/off-category opportunities.
• Post-fix scans run after changed pages are discoverable.

What WhoCites measures

WhoCites does not guess from metadata alone. It runs the category prompts against live AI and search sources, then ties the result back to the pages and signals that explain the miss.

• Weighted 0-100 visibility score.
• Per-engine mention rate and average brand rank.
• Competitor share of voice.
• Citation/source patterns where engines expose links.
• Engine coverage across 7 AI/search sources.

What to fix next

The next fix should improve the source material AI can retrieve, not just the words on a landing page.

• Treat the score as a diagnostic, not a vanity metric.
• Prioritize core prompt misses before adjacent prompts.
• Use the engine breakdown to decide whether the gap is Google/Bing discovery, ChatGPT retrieval, or broader semantic authority.
• Document the before/after score and changed URLs for the next re-scan.

How to measure AI visibility

Measure AI visibility by running the same buying-intent prompts across ChatGPT, Claude, Gemini, Grok, Copilot, Perplexity, and Google AI Overviews, then scoring mention rate, brand rank, share of voice, citation rate, and engine coverage. WhoCites runs that measurement for $49 and returns one 0-100 visibility score.

Mention rate

Mention rate is the percentage of prompts in which the brand is named at all. It is the simplest visibility metric: high mention rate means AI sees the brand; low mention rate means AI does not.

Brand rank

Brand rank is the brand's position in the AI response when mentioned. Lower is better — being the first brand recommended is meaningfully different from being the fourth. Engines treat first-mention bias differently, which is why rank matters alongside mention rate.

Share of voice

Share of voice is the brand's mention count divided by total brand mentions (including competitors) across the scan. It shows whether the brand dominates the AI answer set in its category or is a minority voice next to better-cited competitors.

Citation rate

Citation rate is the percentage of source URLs (that AI engines expose) which point to the brand's own domain versus competitor and third-party domains. It is a direct retrieval signal: high citation rate means the brand's own pages are being pulled into AI answers.

Engine coverage

Engine coverage is how many of the 7 engines mention or cite the brand at all. A brand covered by 6 engines and missing on 1 has a targeted problem; a brand covered by 1 engine and missing on 6 has a systemic problem. Coverage drives the shape of the fix list.

How WhoCites combines the metrics

WhoCites returns one 0-100 visibility score weighted across mention rate and rank per engine, plus the full engine-by-engine breakdown, share of voice, citation summary, and a fix list. The methodology is documented at /methodology so the score is reproducible, not opaque.