iBrandVoice / Cases / AI visibility audit
Fabulous Fido is my own pet-apparel brand. I audited it as a research project, with the same audit I run for clients, before changing anything. Below: what the AI tools did, what to change first, and what is still pending.
Brand mentions in this audit sample; not market share or sales attribution.
The brand was named in 20 of 49 answers to unbranded questions, and in 0 of the 5 Google AI Mode answers checked. Of the answers that named it, 17 of 20 presented it as a suitable pick.
Give Google indexable product pages on the brand's own site and connect the product data to Google Merchant Center. Then ask Google AI Mode the same questions again.
About 9 in 10 answers with text on record (65 of 73) relied on marketplace listings. The brand's own site, then before launch, was cited in 12 of 73.
Align the brand description on Chewy, Amazon, Etsy and the brand's own site, so the listings AI already relies on tell the same story.
A shopper asked for a knitted hat that leaves a French Bulldog's ears uncovered. Gemini's answer listed two Fabulous Fido Cozy Ears beanies and gave them different compositions, citing the same Chewy listing for both.
On the Cozy Ears Beanie
“Material composition: 100% acrylic yarn.”
On the Alpaca-Blend Beanie
“85% acrylic, 15% alpaca wool blend.”
Source cited for both: Chewy, “FABULOUS FIDO Alpaca-Blend Handmade Cozy Ears Beanie”
“Made from an 85% acrylic and 15% alpaca blend”
The September version of the page wasn't archived. In the same answer, Gemini quoted it as 85/15.
The answer recommended the “100% acrylic” hat for sensitive skin. Anyone who follows its source lands on the alpaca blend.
Possible cause, not confirmedThe two versions have nearly the same name, and the answer may have merged their sources. The audit material doesn't establish the cause.
RecommendedState the exact composition of each version and colorway in listing text, on every channel. It lowers the risk of one version's material being attached to another; the follow-up asks the same question to check.
Selected findings shown. Full analysis and implementation priorities are included in the client report. Implementation and follow-up measurement are pending; the follow-up will use the same questions and the same AI tools.
It does not show hidden ranking factors or guaranteed placement in AI answers.
Once several channels change at the same time, it can't attribute a result to one change.
The numbers count brand mentions in one audit sample. They are not market share and not sales attribution.