How AI Sells Adidas: 231 Answers From ChatGPT, Gemini and Perplexity

Adidas appears in 70% of ChatGPT's answers to sneaker questions. Ask Perplexity the same questions and the number drops to 38%. Same brand, same week, same prompts, and a 32-point gap depending on which AI the customer happens to use.

That is the kind of thing you only see when you stop asking AI one question once and start asking the same questions every day, across every model, and keeping the receipts. So that is what we did. This is the first post in our teardown series: one brand, one week of real AI answers, and everything we can learn from them.

What we did

We tracked 11 buying-intent prompts about sneakers and running gear, the questions real shoppers ask, such as "What are the best lightweight racing shoes for faster tempo runs?" and "Looking for cushioned shoes for daily road runs under $150, any suggestions?". Each prompt ran every day for 7 days (July 8 to 15, 2026) across ChatGPT, Gemini and Perplexity, through the real user interfaces rather than API simulations. That produced 231 stored answers. For every answer we recorded which brands appeared, in what order, and which sources the model cited.

One honest caveat before the numbers: this is a one-week window on 11 English-language prompts. It is a teardown of how AI sells one brand right now, not a market study of the entire sneaker industry. As you will see below, the numbers move enough week to week that the snapshot date matters, which is rather the point.

The scoreboard: three models, three different opinions

Across all 231 answers, Adidas appeared in 133, a visibility rate of 58%. The average hides the real story:

If your customers ask ChatGPT, Adidas is close to a default recommendation. If they ask Perplexity, Adidas misses more than 6 answers out of 10. Perplexity is the most citation-driven of the three engines, it builds answers directly from retrieved pages, and as the sources section below shows, the pages that dominate this category are not ones where Adidas wins. A brand can look healthy on one model and invisible on another, and a single "AI visibility" average will hide that completely.

The race: Nike leads twice

Here is the full competitive picture for the week. Visibility is the share of relevant answers mentioning the brand; position is where in the answer the brand appears on average, when it appears at all.

(Also tracked, all below 5%: Reebok, Vans, The North Face, Champion, Converse.)

Nike leads on both metrics that matter, and the second one is easy to miss. Nike does not just appear in more answers than Adidas (66% versus 58%), it appears earlier: average position 1.9 versus 2.8. When an AI assistant lists five options, being the first name and being the third name are different products. AI answers are read top to bottom just like search results, except there is only one answer on the page.

Brooks shows the opposite pattern and it is worth a moment. It appears in only 22% of answers, but when it does appear, it averages position 2.1, nearly as high as Nike. AI treats Brooks as a specialist: rarely raised, taken seriously when raised. Adidas is the inverse, a generalist that gets listed often but rarely first.

Where Adidas wins, and where it disappears

Visibility by topic:

AI sees Adidas as a shoe company. On basketball prompts it is close to omnipresent (40 mentions in 42 answers), and on general running-shoe prompts it appears in 4 answers out of 5. Apparel is another world: barely 1 answer in 3.

The two worst prompts of the week deserve to be quoted in full.

"Where do I look for sustainable running gear for high-mileage training?" produced 21 answers. Adidas appeared in 2 of them.

"Any recommendations for durable trail running shorts that handle mud and abrasion?" produced 21 answers. Adidas appeared in none.

The first one should sting. Adidas has spent years building a sustainability story, from ocean-plastic shoes onward, and it is one of the few sportswear brands with genuine equity on the topic in human minds. In AI answers that equity is simply absent: a 9.5% visibility rate on the exact question the sustainability program was built to answer. Whatever the campaigns achieved in press and awareness, the sources AI reads when a shopper asks about sustainable running gear apparently do not put Adidas on the list.

Who actually shapes the answers

For every answer we also stored which domains the models cited. Across the week, 206 different domains appeared. The top of that list explains almost everything above:

Read that top line again. RunRepeat, a shoe-review site, appears in more than half of all AI answers about this category. Runner's World appears in 4 out of 10. Adidas's own website appears in 3% of answers, and Nike's does barely better at 4.3%. The brands spend the marketing budgets, and a handful of review publications decide what AI says about them.

This is the mechanism behind everything else in this teardown. The model gap? Perplexity leans hardest on retrieved sources, so it amplifies the reviewer economy the most. The sustainability blind spot? If RunRepeat and Runner's World have not crowned Adidas in their sustainable-gear roundups, there is nothing for the models to repeat. In AI search, your brand is what your category's reviewers say it is.

One week of movement

A final table, and the reason snapshots mislead. Week-over-week visibility change:

New Balance lost 15 points of visibility in 7 days. Hoka gained 7. Nothing dramatic happened to either company that week; the answers simply moved, because models update, sources shift, and generation is probabilistic. Any tool that checks your brand once and hands you a score is describing a coin mid-flip. This is why we run every prompt daily and report trends.

Three lessons for any brand

First, your AI visibility is decided by a short list of publications you probably do not monitor. Find the equivalent of RunRepeat for your category (our Sources report exists for exactly this) and win coverage there before rewriting your own product pages.

Second, track models separately. Adidas's 58% average hides a 70/38 split between ChatGPT and Perplexity. If your buyers skew toward one engine, the average is noise.

Third, topic strength does not transfer. The same brand is at 95% in basketball and 35% in apparel. AI does not have one opinion of you; it has one per topic, and each is won or lost separately.

See this for your own brand

Everything in this teardown came from one GetMentioned workspace: 11 prompts, 3 models, daily runs, stored answers with sources. You can see the same picture for your own brand in about five minutes with our free AI visibility report, no signup required: it shows where you appear, who beats you, and which sources drive the answers in your category.

Generate your free AI visibility report

Methodology: 11 buying-intent prompts tracked daily from July 8 to 15, 2026 across ChatGPT, Gemini and Perplexity via their real user interfaces, 231 answers total. Visibility = share of answers mentioning the brand. Position = average rank of the brand within an answer when mentioned. Source share = share of answers citing the domain. Adidas is tracked here as a public demo brand.