How to measure whether AI recommends your brand
Plenty of tools now promise an "AI visibility score". Most hide how the number was made. Here is a simple method you can trust, and the mistakes to avoid.
A method in four rules
- Ask buying questions, never brand questions. "Best self tan for fair skin under £20" tells you something; "Is Brand X good?" does not.
- Repeat each question. The same prompt can name different brands on different runs. Three samples per question is a sensible minimum.
- Always show the denominator. "Mentioned in 3 of 15 answers" can be checked. "Visibility 62" cannot.
- Say where the answers came from. API answers can differ from what a consumer app shows the same person.
Two numbers that are often confused
| Measure | What it means |
|---|---|
| Mention rate | The share of answers that mention your brand at all |
| Share of brand mentions | Your mentions divided by all brand mentions across the answers |
A brand can appear in half the answers yet hold a small share if every answer lists ten competitors. Report both, and never blend them into one score.
What a result can and cannot tell you
A sampled result tells you how often an assistant, asked those questions on that date, mentioned you. It does not tell you what every shopper sees, and "not mentioned in 15 answers" is not proof that AI never recommends you. Re-run the same question set over time and compare like with like.
Revenue Shield offers a free report built on these rules. Run a free AI visibility report.
Frequently asked questions
Why do AI answers about my category keep changing?
Is an API answer the same as what ChatGPT shows a shopper?
See it on your own store — free
Run a free checkout health scan on any live Shopify store. No install, no card. Results in about 8 minutes, with step-by-step video proof.
Scan my store freeor install on Shopify →Rated 5/5 by merchants on the Shopify App Store