Minimalist has a 20.4% Share of Voice in Indian skincare — the highest of any brand we track. Gillette's website scores 78 on AI-readiness. AI mentions it 6 times across 4 platforms. Beardo's website scores 71 and gets 72 mentions.
Three brands, three different stories. And none of them make sense if you're only tracking a single number.
Most brands reduce AI visibility to one composite score. But that number can't tell you why Gillette's technically superior website gets ignored while Beardo dominates AI recommendations. It can't explain why a brand with strong citations might still be losing the category.
The Cited 8 is the framework we built after tracking 202 brands across 8 categories and 4 AI platforms over 6 months. Eight metrics, each mapped to a specific layer of the 3-Layer AI Visibility Stack — so you always know what to measure and where to fix it.
Why a single score isn't enough
A composite AI Visibility Score of 35 could mean any of these:
- Cited on every platform, but only for 3 out of 30 tracked prompts. Narrow visibility.
- Frequent mentions, but AI frames you as "budget" when you've repositioned to premium. Wrong narrative.
- Strong on ChatGPT, zero on Perplexity, Claude, and Gemini. Platform dependency.
- Perfect website, zero third-party coverage. Invisible despite readiness.
Each scenario requires a different fix. A single score can't distinguish between them. The Cited 8 can — because the framework maps every metric to the 3-Layer Stack, so your team knows which layer is the bottleneck.
Layer 1 — Discoverability: Can AI even find you?
If these two metrics are broken, the other six are irrelevant.
Metric 7: Schema & Technical Health Score
The percentage of key pages with proper schema markup (Article, FAQ, Organisation, Product) plus full AI-crawler accessibility. If GPTBot, ClaudeBot, or PerplexityBot can't crawl your site, AI doesn't know you exist. Crawl Radar tests 10 AI crawlers against your site in 30 seconds.
Metric 8: Content Freshness Rate
The percentage of key pages updated in the last 90 days. LLMs deprioritize stale content — industry analysis suggests fresher content gets cited significantly more often across AI platforms. A GEO Score scan checks 15 AI-readiness signals including freshness, no login required.
Fix horizon: Days to weeks. Mechanical fixes — config changes, schema additions, content date updates.
Layer 2 — Citability: Does AI actually use your content?
Want to know how your brand scores on these same metrics?
We'll run 20 prompts across 3 AI platforms and send your report within 24 hours.
This is where the gap between "technically ready" and "actually visible" opens. Across 186 brands, the average Accessibility score was 79.3% but the average Understandability score was just 49.7%. A 29.6-point gap.
Metric 1: AI Citation Rate
The percentage of AI responses across ChatGPT, Gemini, Claude, Perplexity, and Grok that mention your brand for relevant prompts. Build a prompt library of 20-30 prompts your customers actually ask. Run them monthly. Count the percentage that cite you. This is the core metric.
Metric 5: Prompt Coverage
Out of all prompts customers are likely to ask, what percentage include your brand? Citation Rate can be gamed by tracking only prompts you already win. A brand with 80% Citation Rate on 5 prompts and 0% on the other 25 has 13% Prompt Coverage — the real number.
Fix horizon: Weeks to months. Requires content rewrites — definitive statements instead of hedged language, clear entity definitions, structured formatting, and what Google calls non-commodity content.
Layer 3 — Authority: Does AI position you well?
The hardest layer — because it depends on what others say about your brand.
Metric 2: Share of Voice
The percentage of AI responses in your category that mention your brand versus competitors. Minimalist commands 20.4% SoV in skincare — 92 mentions across 4 platforms. The Derma Co sits at 15.3%. But SoV varies wildly by platform — Keka HR scores 100 on Claude but just 15 on Perplexity — an 85-point gap on the same prompts. That's the 5-platform reality.
Metric 4: Brand Sentiment in AI Responses
The qualitative assessment of how AI platforms describe your brand — positive, neutral, or negative — and whether that framing is factually accurate. Being cited with the framing "one of several options" is worse than not being cited when your competitor gets "the market leader." This is where Source Intelligence matters — 95.7% of AI citations on category prompts come from third-party sources, and those sources shape the narrative.
Metric 6: AI Referral Traffic & Conversion Rate
The volume of website visitors arriving from AI platform links plus their conversion rate relative to organic search. AI-referred traffic converts at 5x organic search (14.2% vs 2.8%, per Superprompt's analysis of 12.3 million visits) — making every improvement in the other 7 metrics directly translatable to revenue.
Metric 3: AI Visibility Score
A weighted 0-100 score combining Citation Rate, mention frequency, Sentiment, and platform coverage across all tracked AI platforms. Use it for executive reporting and month-over-month trending. Use the individual metrics for diagnosis.
Fix horizon: Months to quarters. Requires third-party coverage — press, reviews, expert endorsements — that AI platforms trust.
The full Cited 8 at a glance
| # | Metric | What It Measures | Stack Layer | Fix Horizon |
|---|---|---|---|---|
| 1 | AI Citation Rate | % of AI responses citing your brand | Layer 2 | Weeks-months |
| 2 | Share of Voice | Your brand vs competitors in AI responses | Layer 3 | Months-quarters |
| 3 | AI Visibility Score | Composite 0-100 of citation, sentiment, coverage | Layer 2 + 3 | Varies |
| 4 | Brand Sentiment in AI Responses | How AI describes you (positive/neutral/negative) | Layer 3 | Months-quarters |
| 5 | Prompt Coverage | % of customer prompts where your brand appears | Layer 2 | Weeks-months |
| 6 | AI Referral Traffic & Conversion Rate | Visits + conversion rate from AI platforms | Layer 2 + 3 | Months-quarters |
| 7 | Schema & Technical Health Score | Schema markup + AI-crawler accessibility | Layer 1 | Days-weeks |
| 8 | Content Freshness Rate | % of key pages updated in last 90 days | Layer 1 | Days-weeks |
How the metrics interact
Individual metrics are useful. Metric combinations are diagnostic. Here are the patterns we see across 202 brands.
High Citation Rate + Low Prompt Coverage = Narrow visibility. You're winning the prompts you track but missing the ones customers actually ask. Most common false positive in AI visibility measurement.
High Share of Voice + Low Sentiment = Risky positioning. AI mentions you more than competitors, but frames you poorly — often as the "budget" option. Volume masks the damage until you check Metric 4.
High GEO Score + Low AI Visibility Score = Invisible Ready. We covered this pattern in the 3-Layer Stack article. Gillette scores 78 on GEO Score but just 6 on AI Visibility. Beardo scores 71 and 72 respectively. Eighteen brands fall into this quadrant. Layer 1 is solved. Layers 2 and 3 are missing.
Strong on one platform + weak on others = Concentration risk. Carlton scores 33.3 on ChatGPT and 0 on both Gemini and Perplexity. Each AI platform has a distinct retrieval pipeline — what Google calls query fan-out works differently on every engine. Diversification isn't optional.
The diagnostic path: finding your bottleneck
The Stack tells you where to look. The Cited 8 tells you what to measure.
Start at Layer 1. Run GEO Score and Crawl Radar. If Schema & Technical Health is below 60% or Content Freshness is below 50%, fix the foundation first.
Move to Layer 2. Check AI Citation Rate and Prompt Coverage. If Citation Rate is above zero but Coverage is narrow, you have a content breadth problem. If both are low despite clean Layer 1, your content isn't structured for AI extraction.
Then Layer 3. Check Share of Voice, Sentiment, and Referral Traffic. Low SoV means weak third-party footprint. Negative Sentiment means the narrative needs correction. Growing referral traffic with poor conversion suggests a gap between what AI promises and what your landing page delivers.
| Your Pattern | Likely Bottleneck | First Fix |
|---|---|---|
| Low Metrics 7 + 8 | Layer 1 | Fix crawl access, add schema, update stale pages |
| Clean Layer 1, low Metrics 1 + 5 | Layer 2 | Rewrite content — definitive, structured, original |
| Decent Metrics 1 + 5, low Metrics 2 + 4 | Layer 3 | Build third-party coverage — reviews, press, expert mentions |
| High Metric 1, low Metric 5 | Layer 2 (breadth) | Expand prompt library, create content for uncovered intents |
| High Metric 2, low Metric 4 | Layer 3 (narrative) | Audit AI responses, correct framing via source outreach |
| Strong single-platform, weak others | Diversification | Audit per-platform citation behavior, vary content formats |
How Cited tracks all 8
Metrics 7 and 8 are covered by free tools. GEO Score scans 15 AI-readiness signals in 30 seconds. Crawl Radar tests 10 AI crawlers. No login, no cost.
Metrics 1 through 6 require tracking AI responses over time. The Cited platform runs your prompt library daily across ChatGPT, Gemini, Perplexity, Claude, and Grok — parsing each response for citations, sentiment, positioning, and competitive SoV. Source Intelligence maps which third-party sources drive your citations. Narrative Intelligence shows the gap between your desired positioning and AI's actual framing. Task OS converts every insight into a verified, prioritized action.
But the Cited 8 is an open framework. Track what you can with the tools you have. The measurement matters more than which tool does it.
How to audit your brand across all 8 metrics
The fastest way to find out where you stand across all 8 metrics: request a free AI visibility audit. It tests 20 prompts across ChatGPT, Gemini, and Perplexity — covering Metrics 1 through 6 — and shows you exactly which layer is your bottleneck. Start with GEO Score for Metrics 7 and 8, then layer on the audit for the complete picture. Delivered within 24 hours, no commitment required.