Scoring Basics

A reference for every metric in GrackerAI: what each number means, how it’s calculated, and where to find it in the dashboard.


Core KPIs

Eight cards make up the KPI row at the top of the Overview tab on Insights → Visibility: Visibility, Presence Rate, SOV, Citation Rate, Sentiment, Avg Position, Site Health, and Hero Rate. Hover the help icon on any card for its definition.

1. Visibility Score

Your brand’s overall standing in AI answers, shown as a number from 0 to 100 (for example 5.2 - it is a score, not a percent). Each AI response gets its own score from four signals, and the card shows the average across responses in your date range:

SignalWeightHow it’s scored per response
Presence30%100 if your brand is mentioned, 0 if not
Prominence30%100 if you’re mentioned in the first third of the answer, 66 in the middle third, 33 in the last third, 0 if not mentioned
Sentiment20%Your sentiment (-100 to +100) rescaled to 0-100; 0 if not mentioned
Citation20%100 if the response cites a page on your own domain, 0 if not
If you’re not getting mentioned at all, fixing that moves the needle faster than anything else. See Visibility Score for benchmarks and improvement steps.

2. Presence Rate

The share of tracked AI responses that mention your brand at all - regardless of position or tone. A quick read on raw coverage: are you even in the conversation?

If Presence Rate is low, every other metric becomes unreliable. Fix content gaps here before optimising anything else.

Find it: Overview tab KPI row and Presence Rate chart, or the Presence Rate column on Actions → Prompts Visibility


3. Sentiment

The average tone of your brand’s mentions, from -100 (very negative) through 0 (neutral) to +100 (very positive). It measures how favourably AI describes you, not how often you appear. Calculated only on your brand’s mentions - a response with no mention doesn’t count.

The KPI card shows the number. Where the dashboard shows a word instead (for example in a prompt’s detail view), it uses these bands:

RangeLabel
Above +20Positive - AI describes you favourably
0 to +20Neutral - no strong positive or negative framing
Below 0Negative - AI describes you unfavourably
You can have high Presence with neutral or negative Sentiment - they’re independent. Check both before drawing conclusions about how AI is representing your brand.

Find it: KPI row, or Sentiment & Position tab → Sentiment Drift and Sentiment Distribution


4. Share of Voice (SOV)

Your mentions divided by all mentions of your brand plus your tracked competitors, counting every mention in every response. 100% means only your brand was named; 50% means you got half of all mentions. Archived competitors don’t count.

Find it: KPI row (labelled SOV), Overview tab → Share of Voice (100% Stacked), and Competitors tab → Share of Voice - Totals


5. Citation Rate

The share of AI responses that cite at least one page on your own domain. Higher means AI is not just mentioning you - it’s pointing readers to your site as a source.

Find it: KPI row, or the Citations tab for the full breakdown


6. Avg Position

Where your brand typically appears in a response, shown as a percentile - lower is better (0% = the very start of the answer, 100% = the very end). Averaged across all your mentions in the selected window.

Find it: KPI row, or Competitors tab → Position Battle (Avg Position %)


7. Site Health

Average AEO Audit score (0–100) across your brand’s audited pages - how well your pages are optimised for AI answer engines. Each page’s own score is its Page Health. The card reads “No audited pages” until you audit at least one page. Unlike the metrics above, it’s point-in-time (not tied to your date range). If your site blocks AI crawlers, Site Health is capped at 40.

Find it: KPI row; run audits from Audit → AEO Audit


8. Hero Rate

The share of your brand’s mentions that land in the first third of the AI answer. A high Hero Rate means AI tends to lead with you.

Find it: KPI row, or Sentiment & Position tab → Position Distribution (the Hero tier)


Competitive Metrics

1. Win Rate vs Competitors

Of the responses that mention you or a given competitor, the share that mention you: (Both + Brand Only) ÷ (Both + Brand Only + Competitor Only). Responses that mention neither of you are left out. Tells you which competitors you hold your own against - and which ones are effectively shutting you out.

Find it: Competitors tab → Win Rate vs Competitors · Head-to-Head


2. Mention Position

How early your brand appears in an AI answer compared with competitors. Being named first signals stronger authority to both the reader and the model. The Position Battle table shows average and best position percentile, plus mention count, for you and each tracked competitor.

Find it: Competitors tab → Mention Order vs Competitors · Position Battle


3. Co-occurrences

Pairs of brands the model names together in one response, and how it relates them (compared, alternative to, grouped with, and so on). Reveals the competitive set AI perceives you to be part of. If you’re always grouped with one dominant competitor, that competitor is setting the frame for your category - study what they’re being cited for.

Find it: open any response (Data → Responses) → Co-occur tab


Opportunity & Gap Metrics

1. Lost Prompts

Prompts where your brand was absent across all runs in the window, but at least one competitor was mentioned. These are your highest-value content opportunities - a competitor has already earned AI’s trust for that query and you haven’t yet.

Lost Prompts is the fastest way to find your next piece of content. Pick the top 3, read the actual AI responses, and write content that addresses why your competitor is being cited there instead of you.

Find it: Prompts tab → Lost Prompts


2. Weakest and Strongest Prompts

Your bottom 10 and top 10 prompts by visibility score. Start with Weakest whenever your overall score drops, so you know which prompts are dragging it down before taking action.

Find it: Prompts tab → Weakest Prompts / Strongest Prompts


3. Citation Mix

How the citations in your responses split between your own domain, competitor domains, and neutral third-party sites.

Find it: Citations tab → Citation Mix


4. Citation Source Types

What kind of sites AI cites - news, blogs, forums, review sites, documentation, and more. Shows which source types earn citations in your category so you know where to invest next. See the Glossary for the full list of types.

Find it: Citations tab → Citation Source Types


5. Discovered Entities

Brands the model has started mentioning in your monitored responses that aren’t in your competitor list yet. New entrants typically appear in AI responses before they show up in your sales pipeline. Review regularly so your tracking stays complete.

Find it: Data → Discovered Entities - promote to tracked competitors from there. See Discovered Entities.


Next Steps

Last updated Nov 12, 2026 View .md

Do not let AI keep
recommending someone else

Start your 7-day free trial and get your AI Visibility Score in about a minute. See exactly where you stand, where competitors are beating you, and the ranked fixes to get into the answer. Cancel anytime before the trial ends.

7-day free trial. Cancel anytime before it ends. Trusted by 500+ B2B SaaS teams.

Partnered with
Powered by