The metric
The AI Says Score™
One number between 0 and 100 describing how well AI platforms find, understand and recommend your business. Here is exactly how it is built — including what it cannot tell you.
Illustrative dial. The band label, not the colour, carries the rating.
What the number represents
The AI Says Score answers a single question: if a customer in your category asked an AI assistant for help right now, how likely is it that your business would be named, described correctly and recommended?
It is a composite. Seven components are scored independently on a 0–100 scale, each one is weighted by how much it contributes to being part of an answer, and the weighted average becomes your score. Every component is shown separately, because the composite tells you where you stand and the components tell you what to do.
Crucially, the score does not attempt to reverse-engineer any platform's internal ranking. Nobody outside those companies knows their selection rules, and they change. What can be measured is the observable outcome — whether you appear, how you are described, whether you are recommended, and whether the answer cites you.
Scoring architecture
The seven components and their weights
Weights are configuration, not code. They are versioned so that a change to the methodology never silently rewrites your history.
| Component | Weight | What it measures | Source |
|---|---|---|---|
| AI Visibility | 25% | How often your business appears when customers ask AI questions in your category. | Observed or estimated |
| AI Recommendation | 20% | How often AI actively recommends you, rather than simply mentioning you. | Observed or estimated |
| Brand Accuracy | 15% | Whether what AI says about your business is correct and complete. | Observed or estimated |
| Citations & Authority | 15% | Whether AI answers cite your website and other credible sources about you. | Observed or estimated |
| AEO / Website Readiness | 10% | Whether your website is structured so AI systems can read and quote it. | Measured |
| Competitive Share | 10% | Your share of AI answers compared with the competitors you named. | Observed only |
| AI Sentiment | 5% | The tone AI uses when it describes your business. | Observed only |
| Total | 100% | Components with no data are excluded and their weight redistributed. | |
Rating bands
What each score means in practice
- 85–100
Excellent
AI knows you, describes you accurately and recommends you consistently.
- 70–84
Strong
AI knows your business well, with clear room to be recommended more often.
- 55–69
Moderate
AI has a partial picture of your business and surfaces you inconsistently.
- 40–54
Weak
AI struggles to identify you and rarely includes you in answers.
- 0–39
Critical
AI has little or no usable understanding of your business.
Data confidence is separate — and stays separate
A score of 72 built from 80 observations across four platforms and a score of 72 estimated from a website alone are not the same claim. Folding confidence into the score would hide that difference behind a single number.
So AI Says reports confidence as its own High / Medium / Low rating, derived from how many answers were observed, across how many platforms, how recently, and whether your website could be retrieved. Confidence reflects the amount and quality of data available for this assessment. It is shown separately and is never folded into the AI Says Score.
Observed, estimated, unavailable
- Observed
- Computed from recorded AI answers, or measured directly from your website in the case of AEO Readiness.
- Estimated
- Approximated from measured website and entity signals using published heuristics, capped below the maximum, and always labelled.
- Unavailable
- Competitive Share and AI Sentiment cannot be honestly inferred from a website, so without observations they are excluded rather than guessed.
Limitations
What this score cannot tell you
A measurement product that only lists its strengths is a marketing product.
It is not deterministic
Ask the same question twice and a generative system can answer differently. Location, personalisation, model version and phrasing all move the result. Treat a single reading as a sample, not a fact.
It is not a platform ranking
No AI provider publishes a business ranking, and AI Says does not claim access to one. The score measures outcomes visible in answers, which is a different and more honest thing.
It is not a guarantee
Acting on every recommendation makes you more findable, more quotable and better corroborated. It does not oblige any platform to name you, and nobody can promise that it will.
AI Says Score™ is a proprietary visibility assessment developed by AI Says. AI-generated answers may vary by platform, model, location, personalisation, prompt wording and time.
AI Says is not affiliated with or endorsed by OpenAI, Google, Anthropic, Perplexity or other AI/search providers. Platform names are used descriptively to identify the services being assessed.
Methodology version: AI Says Scoring Methodology v1.0. Read the full disclaimer.
Questions
Questions about the score
No. It is a proprietary visibility assessment developed by AI Says. No AI or search provider publishes a ranking for businesses, and any product claiming to report one is inferring it. The score measures observable outcomes — whether you are named, recommended, described correctly and cited — not a platform-internal rank.
Because competitive share is a consequence, not a cause. A business that is invisible and poorly described will lose share automatically. Weighting the causes higher means the score tells you what to change rather than restating the outcome.
The component is excluded and its weight is redistributed proportionally across the components that do have data. The score is never padded with a neutral default, and the excluded component is shown as "No data" rather than as a zero.
When no AI platform integration is connected, five of the seven components are approximated from measured signals on your own website using the published heuristics in the scoring engine, and are capped well below 100. They are labelled Estimated, the score is labelled Preliminary, and data confidence drops to Low.
Website-derived components will, because they are measured. Components derived from AI answers will vary, because generative systems are not deterministic. That variance is why confidence is reported separately and why monitoring over time is more meaningful than any single reading.
It will, as AI search matures. Every score records the methodology version it was calculated under, so comparisons across versions are never made silently. The current version is AI Says Scoring Methodology v1.0.
See your own AI Says Score
Enter your website and get your AI Says Score. No credit card required.