Share of AI answers (AI mention rate)
In short: Share of AI answers (also called mention rate or AI visibility score) is the percentage of answers that mention your brand or store when a fixed set of questions is asked to AI assistants. It is only meaningful if the questions stay the same over time and you compare against a baseline.
Definition
Share of AI answers is how often AI assistants mention you when asked a set of questions. If you ask 30 questions and 6 answers name your store, your share is 20%.
Tools use different names for it: mention rate, AI visibility score, share of voice in AI. Wix's AI Visibility Overview, for example, defines its score as "the percentage of questions in which your site was mentioned out of the total number of questions generated". There is no single industry standard, so numbers from two tools don't compare directly.
What makes the number honest
- Fixed questions. Ask the same questions every time. If the questions change, the number changes for that reason alone.
- Real shopper questions. "Best linen apron for a home baker" is a shopper question. "Is Example Studio good?" is a brand question, and of course it mentions you. Keep the two apart.
- A baseline. Measure before you change anything. Otherwise you can't tell what changed.
- Honest limits. A rise after your changes is a good sign, not proof. The AI itself changes too (new models, new sources), and a before and after can't tell the two apart.
- Clear mention rules. Decide what counts: the store's link or web address, or only the name? A name that the question itself contained should not count.
- Several runs. AI answers vary. One week is noise; look at a trend.
Why a store owner should care
It is the most direct way to see whether AI assistants pick your store when shoppers ask. But it is easy to fool yourself with a number that moves for reasons nobody can check.
How flareOps measures it
flareOps asks ChatGPT and Perplexity about 30 shopper questions in your store's language, plus 2 brand questions, every week. A baseline runs before any change, and every week is compared with it. A store name that the question already contains does not count as a mention. The report shows each question, each answer, and whether you were found. See our method.
A real example
Week 1 (baseline): 3 of 30 shopper questions mention you, 10%. You fix brand and description on your products. Weeks 5 to 8: between 7 and 9 of 30, about 27%. A rise that holds for several weeks is worth more than one good week. It still doesn't prove your fixes did it, and it is not a guarantee for the future.
How to act on it
Questions people ask
Why does my score move when I changed nothing?
AI answers vary from run to run, and the AI models themselves change. That is why one test means little. Look at the trend over several weeks, compared with your baseline.
Should questions include my store name?
Mostly not. A question that already names your store will almost always mention it. Count those separately as brand questions. Generic shopper questions show whether AI picks you on its own.
Is a higher share the same as more sales?
Not always. Check AI referral traffic and orders as well. Mentions show visibility; visits and sales show the result.
Sources
- Wix Help Center: AI Visibility Overview (example definition of an AI visibility score)
- Aggarwal et al., GEO: Generative Engine Optimization (arXiv, 2023)
Sources checked on 4 November 2026.