Your Visibility Score is a comprehensive 0–100 metric that quantifies how prominently your brand appears across AI engines. It’s the single most important number in your AEO strategy.
Not ready to set up tracking yet? The free AI visibility score checker gives you a one-off read on a single domain (a live two-engine, two-prompt sample) before you decide how to monitor it properly.
Composite scores like this exist because generative answers have no ranked list to read a position off. The research literature hit the same wall: GEO: Generative Engine Optimization had to define its own impression metrics (weighting a source by how prominently it features in the generated text) rather than reuse the click-and-rank measures search handed down. The weighting below is the same idea applied to a brand rather than a URL.
What Does the Score Measure?
Three components, weighted 50 / 35 / 15. Nothing else feeds the score. Citations, competitor mentions and claims are recorded on the same answer and drive the Citations, Competitors and Alerts pages. But they are not score inputs, so optimising specifically to raise a “citation rate” will not move this number.
1. Mention rate (50%)
The share of delivered answers that name your brand at all. If an engine answered 20 of your prompts and named you in 10, that is 0.5.
The denominator is answers that came back successfully. An answer an engine failed to return is excluded rather than counted as a miss, which is also why a provider outage costs you nothing and never depresses the score.
2. Prominence (35%)
Where you sit among the brands named in the answer, converted to a 0–1 factor:
| Position among named brands | Factor |
|---|---|
| 1st | 1.0 |
| 2nd | 0.8 |
| 3rd | 0.6 |
| 4th–7th | 0.4 |
| 8th or later, or named with no discernible order | 0.25 |
These are multipliers, not points: there is no running total to accumulate. An unordered mention scores the same as the tail of a long list: the engine named you, but not as a ranked recommendation. It still counts fully towards mention rate.
3. Sentiment (15%)
The tone of the answers that mention you: positive scores 1, neutral 0.6, negative 0.
Prominence and sentiment are averaged over mentioning answers only. Mention rate uses every delivered answer; the other two use only the subset where you were named. That asymmetry is deliberate (averaging prominence over answers that never mentioned you would double-count the same absence), and it is the single most important thing to understand about how the components interact. A brand named in 1 answer out of 20, first and positively, keeps a perfect prominence and sentiment score on a sample of one.
The one exception: with zero mentions the sentiment term is forced to 0 rather than defaulting to neutral, so a brand no engine ever names scores 0 rather than inheriting a floor of 9.
How the Score is Calculated
For one engine that delivered 20 answers and named the brand in 10 of them (two of those first, two second, two third, one fourth, one sixth, and two naming it with no readable ordering; seven positive, two neutral, one negative):
| Component | Value | Weight | Contribution |
|---|---|---|---|
| Mention rate | 10 / 20 = 0.50 | 50% | 25.00 |
| Prominence | (1.0 + 1.0 + 0.8 + 0.8 + 0.6 + 0.6 + 0.4 + 0.4 + 0.25 + 0.25) ÷ 10 = 0.61 | 35% | 21.35 |
| Sentiment | (7 × 1 + 2 × 0.6 + 1 × 0) ÷ 10 = 0.82 | 15% | 12.30 |
| Engine score | 59 |
The result is rounded to a whole number: the score is always an integer, so a fractional figure is not something the product can produce.
From per-engine scores to the one you see
Your headline score is the answer-count-weighted mean of the per-engine scores, not a plain average of them. An engine that delivered one answer counts one twentieth as much as an engine that delivered twenty. That matters because a plain mean of means lets a single flaky engine that answered once move the headline number by tens of points, and the number is frozen onto the scan when it is written, so it would carry that distortion into the trend line and every downstream comparison permanently.
Score Benchmarks
0–20: Minimal. Minimal visibility; rarely mentioned. Significant opportunity for growth. Focus on content quality and authority.
21–40: Emerging. Inconsistent mentions, mostly listed. Some competitive presence established. Continue building authority.
41–60: Competitive. Regular mentions, mixed positioning. Competitive positioning established. Opportunity to move toward first mentions.
61–80: Strong. Frequent mentions with prominent positions. Leading in category for AI visibility. Focus on maintaining and protecting position.
81–100: Dominant. Category leader with high first-mention rate. Your brand dominates in AI responses.
Three anchor points make those bands concrete, because mention rate carries half the weight and therefore dominates which band you land in:
- Named in every answer, mid-list, neutrally → 73. Ubiquity alone gets you into the fourth band without a single first mention.
- Named in half the answers, around third place, neutrally → 55.
- Named in no answer → 0, with no partial credit from the sentiment default.
The corollary is worth internalising before you read your own number: a brand at a 10% mention rate cannot exceed 55 even scoring perfectly on both other components, and a brand named in every answer cannot fall below 68 however far down the list it sits, provided the mentions are at worst neutral. With uniformly negative sentiment and bottom-of-list placement the floor is closer to 59. Read the score beside your mention rate, never alone.
What Influences Your Score?
The score is computed purely from the text of the answers engines returned; nothing about your site is measured directly. So the factors below are not score inputs; they are the levers that change whether an engine names you, which is what the score measures. Expect them to move the number indirectly and on the engines’ refresh timescale, not on your deploy’s.
Content Factors
- Freshness: Recently updated content ranks higher
- Depth: Comprehensive, authoritative content
- Uniqueness: Original research and perspectives
- Structure: Well-organized, scannable content
Authority Factors
- Backlink Profile: Quality and quantity of inbound links
- Domain Authority: Overall trust and credibility
- Brand Mentions: Citations across the web
- Industry Recognition: Awards, certifications, expert positioning
Technical Factors
- Website Performance: Speed, uptime, accessibility
- Mobile Optimization: Responsive design
- Schema Markup: Proper structured data implementation
- Indexation: Proper crawlability and indexing
External Factors
- Industry Trends: Seasonal or market-driven changes
- Competitor Activity: Competitive positioning
- AI Model Updates: Changes to how engines work
- Training Data Changes: Shifts in AI training datasets
Improving Your Score
Month 1: Foundation. Audit your current visibility. Identify core queries where you should appear. Begin monitoring baseline metrics.
Month 2: Content. Publish high-quality, authoritative content. Update existing content for relevance. Focus on comprehensive coverage of key topics.
Month 3: Authority. Build backlinks from relevant sources. Increase brand mentions and citations. Pursue industry recognition.
Ongoing: Optimization. Track mention positioning and sentiment. Respond to competitive threats. Continuously refine content and authority strategy.
What a Trend Looks Like
We do not publish an expected improvement curve, because we have no basis for one that would survive contact with your category. The honest shape of the answer is “it depends on how contested your queries are and how much authority you start with”, and a number attached to that would be invented.
What is worth knowing is why the line moves the way it does. Your prompts are re-run on your plan’s cadence, so the score is a series of discrete readings rather than a continuous measure, and two consecutive scans of an unchanged site will rarely be identical: generative answers vary run to run. Movement of a few points between scans is the measurement, not the market. Judge the direction over several scans, and read a large simultaneous shift across many prompts as a likely model update rather than as something you did; see why queries return different results.
Frequently Asked Questions
What is a visibility score?
A visibility score is a composite metric that summarizes how present your brand is across AI engines. Ours blends three signals into a single trackable number: how often you’re mentioned (50%), how prominently you’re placed among the brands named in the answer (35%), and the sentiment of those mentions (15%). Citations are recorded against the same answers but are not part of the blend.
What is a good AI visibility score?
There’s no universal threshold: what matters is your score relative to competitors and your own trend over time. A rising score and a higher share of voice than rivals in your category matter more than any absolute number. For benchmarks by stage, see what is a good visibility score.
How can I improve my visibility score?
Build topical authority with deep, citable content, strengthen your entity signals, earn corroborating mentions from reputable sources, and structure content so AI engines can extract and quote it. Then re-measure and double down on what moves the score.
How often does the visibility score change?
It shifts as AI engines refresh retrieved content (which can happen within days) and as models update (which is slower). You see that as a new reading each time your prompts are re-run: daily on every paid plan (weekly on free), and on demand whenever you run a scan yourself. Tracking it across scans, rather than as a one-off snapshot, is what reveals the trend.