Case study · Pet supplements · United States
We looked at what AI assistants tell dog owners who ask for help with a daily supplement, and what those assistants say about one US brand in the category.
01What they believed
They believed one rival was taking their buyers' attention, a brand they saw everywhere they sold.
They believed their buyers came to them for three reasons in roughly equal measure.
And they believed that strong reviews on the big marketplaces meant the assistants knew them well.
02What surprised them
The rival they worried about appeared in 2% of Google's AI answers, and too rarely in the assistants to count. A different group of brands holds that ground. The leader was named in 27% of answers, which still leaves it out of 73%. Nobody owns this shelf yet.
The biggest single cause of negative answers about them was some of their own wording, on their own pages.
Even the leader is missing from most answers
Share of recommendation answers that named the category leader
Source: study data, August 2026
03What we found
Six findings, each with the exhibit that shows it.
Finding 1
When an owner asked for a recommendation, they were in 0% of the answers. That held in every assistant, including the ones that searched the web first. Searching did not help.
Who gets named when owners ask for a recommendation
Share of recommendation answers naming each brand
Source: study data, August 2026
24%
of the claims we checked against their own catalogue were wrong
Where the assistants lacked facts, they filled the gap.
Finding 2
Google's AI answers were the only place they appeared on their own, in 4% of them.
Google's AI answers: a thin opening
Share of Google's AI answers naming each brand
Source: study data, August 2026
Finding 3
Asked about by name, the assistants knew them, and 58% of what they said about the product was negative. Two of the brands most often recommended drew 2% and 3% negative. 56% of all the criticism came back to two causes. One was proof: the assistants could not find evidence for gruampdkek, and repeated the doubts of outside sources instead. The other was a claim on the brand's own pages, oylamxgojj, which the assistants would not repeat.
The same assistants, very different verdicts
What the assistants said about each product, positive against negative
Source: study data, August 2026
Two causes carry more than half of the criticism
Proof the assistants could not find
Outside doubts about yqartzbopo, repeated by the assistants.
Claims on their own pages
spyqlzgdqq, which the assistants would not repeat.
Finding 4
The assistants learned about them from the marketplaces that sell them. Their own website was barely read. Their competitors' own sites were cited directly.
Where the assistants read
Finding 5
Where the assistants lacked facts, they filled the gap. 24% of the claims we checked against the brand's own catalogue were wrong.
of checked claims were wrong
Checked against the brand's own catalogue
of answers about them sent the owner to a competitor
Asked about the brand by name
Finding 6
Asked about them, the assistants sometimes sent the owner elsewhere: to a competitor in 6% of answers.
Their buyers mostly came for one reason. What owners wrote in reviews put one need far ahead of the other two, so the study asked the questions those owners actually ask.
04What they are doing
Our studies come with a detailed, private action plan that tells this brand what to do, what to change, what to improve and where, with KPIs to monitor and measure the impact, so they can flip the page to their side.
Your brand
Every brand's picture is different. A free assessment shows you yours.