• AI visibility

Guide · AI search

How do sportswear brands win the sale when shoppers ask AI what to train in?

By being the product an assistant names when a shopper describes an activity, a constraint and a budget, and by making that product easy to buy wherever the answer points. Sportswear questions are unusually specific: anti-chafe shorts for a half marathon, a high-support bra for running, a soccer kit for a hot summer. The brands that answer those specifics clearly, on their own pages and through retailers and reviewers, give AI something concrete to recommend.

By Underneath Team Guide Updated

The short version

  1. Participation is at a record: the Sports & Fitness Industry Association (opens in a new tab) counted 250 million active Americans in 2025 across 126 activities, with team sports above 90 million participants for the first time.
  2. Owned digital sales are diverging across the leaders: Nike’s (opens in a new tab) brand digital sales fell 12% in fiscal 2026, lululemon’s (opens in a new tab) ecommerce fell 6% while still making up 39% of revenue, and adidas (opens in a new tab) grew ecommerce 27% in its World Cup quarter.
  3. Google now documents sports-gear questions with constraints, such as “lightweight hydration vests under $80,” and says searches like “run club” and “how to train for a marathon” hit all-time highs this year (Google (opens in a new tab)).
  4. The sale can now finish inside the chat: JD Sports (opens in a new tab) set up one-click purchases on AI platforms such as ChatGPT and Gemini for US shoppers.
  5. A famous name helps only when products look the same: in a controlled test, product details explained 82.4% of AI rankings and brand identity 1.2% (Chu and Hou (opens in a new tab)).

Who buys sportswear, and where does the money come in today?

Active people buy by activity and season, through a mix of brand sites, retailers and marketplaces.

The buyer is whoever is about to do something: a first-time marathoner, a parent outfitting a youth team, a pilates regular, a pickleball convert. The SFIA’s 2026 report says pickleball has been the fastest-growing sport for five years running, and its members, more than 700 brands, manufacturers, retailers and governing bodies, generate $150 billion in domestic sales. Each new activity brings a short list of needs: what to wear, what to carry, what to put on your feet.

How that demand turns into revenue varies a lot by brand. Nike earned $46.4 billion in fiscal 2026, of which wholesale was $27.5 billion and its own direct business $17.7 billion, and it spent $4.8 billion on demand creation. Adidas reported record second-quarter sales of €6.7 billion, with direct-to-consumer sales up 25% and apparel up 35%; it also spent an additional €212 million on marketing and World Cup activations. Lululemon’s ecommerce brought in about $900 million in its second quarter, and its interim co-CEO said the company has redesigned its homepage and category pages and will update its product pages next.

Two lessons follow for executives. First, sportswear brands already spend heavily to create demand, so an assistant naming the product for free at the moment of need is valuable. Second, a large share of sales runs through retailers, so the brand does not always own the page where the purchase happens. AI answers can point to either.

Why do sports moments send shoppers to AI assistants?

Because a race, season or new sport creates many questions, and AI answers them together.

Google’s September 2026 post on race preparation shows the pattern from the platform’s side. It says running-related searches are “spiking,” with “run club,” “how to choose running shoes” and “how to train for a marathon” at all-time highs. It then suggests using AI Mode to build a training plan and to shop for gear that fits constraints, such as “road shoes for wide feet, lightweight hydration vests under $80, or anti-chafing apparel,” drawing on what it calls a Shopping Graph of over 60 billion product listings, with side-by-side comparisons and local availability.

Events work the same way at a larger scale. Adidas credits its World Cup collections and campaigns, plus its football and running categories, for a 39% rise in its performance division. Our inference is that event moments concentrate shopping questions into a few weeks, which is when being named matters most.

Shoppers are bringing more of these questions to AI in general. In Adobe’s retail data (opens in a new tab), reported by TechCrunch, AI-driven revenue per visit to US retail sites was 37% higher than non-AI traffic in March 2026, a reversal from a year earlier. That figure covers all of retail; nobody has yet published a sportswear-only number.

Which shopping prompts carry the most purchase intent in sportswear?

Prompts that combine an activity, a performance attribute, a fit detail and a price.

We wrote the examples below to illustrate high-intent sportswear prompts; they are not observed data:

  • Attribute plus activity: “Squat-proof leggings with a phone pocket for heavy lifting, under $100.”
  • Problem to solve: “Running shorts that don’t chafe on long runs in humid weather.”
  • Fit and support: “High-support sports bra for running that actually fits a larger cup size.”
  • New sport: “Do I need pickleball shoes, or are tennis shoes fine?”
  • Team and bulk: “Matching training tops for a youth soccer club of 20 players, with numbers.”
  • Event: “Breathable football shirt to wear to World Cup games in summer heat.”
  • Buy now: “Black running shoes under $150, good for daily training.”

The last example is the one The Next Web used to describe JD Sports’ AI shopping flow (opens in a new tab): a shopper refines the request, compares options and pays in the same chat. Prompts like these carry more purchase intent than a broad “best activewear brands” question, because the shopper has already decided what they need and is choosing which product delivers it. For the difference between your brand being known and being chosen for a plain category question, see our article on why well-known brands miss AI recommendations.

How does an AI answer turn into a sportswear sale, and who books it?

The answer names products, then sends the shopper to a brand site, a retailer, or a checkout inside the assistant.

There are now three endings to the same conversation.

On your own site. The shopper clicks through from a product result. This is the highest-margin ending, and it depends on your product pages saying what the shopper asked about. Adobe found that around 34% of retail product pages cannot be properly accessed by AI, which is a risk for brands whose product details sit in images or scripts.

At a retailer. OpenAI documents (opens in a new tab) that when a shopper opens a product, ChatGPT may list several merchants, ranked “based on factors like availability, price, quality, and whether they are the maker or primary seller of that item.” A retailer with complete stock and price data can win that click even for your product. For the marketplace side of that click, see how online marketplaces win buyers and sellers.

Inside the assistant. OpenAI’s help page says that for some eligible products and merchants, ChatGPT may show an Instant Checkout option so the shopper can pay without leaving. OpenAI scaled the feature back in March 2026, according to FashionUnited (opens in a new tab), though that help page still describes it. Google’s (opens in a new tab) agentic checkout lets a shopper track a price, then tap “buy for me” to complete the purchase on the merchant’s site through Google Pay. JD Sports, which calls North America its largest market, chose to make its catalog purchasable in AI tools; its chief executive said the aim is to let customers using AI “find and transact with JD quicker.”

For a brand, the strategic question is less “will AI send traffic?” and more “when AI recommends our product, which of these endings happens, and at what margin?” We infer that brands with clean product feeds and strong retailer data will capture more of the second and third endings even when they do not win the first.

Does a famous logo decide which sportswear AI recommends?

Only when products look alike; clearly stated advantages decide the rest, and answers vary from run to run.

Observed in a controlled study. Chu and Hou (opens in a new tab) gave three AI models lists of products in which one brand was real and the rest invented. When every product had identical specs, the real brand was recommended in all 670 valid trials. But when an invented brand had even the smallest advantage in rating, price or reviews, its win rate jumped from 3.6–6.0% to 64–80%. Product details explained 82.4% of the ranking, brand identity 1.2%. For a challenger activewear label, that is encouraging: a better-documented product can beat a household name. For a market leader, it is a warning not to rely on the logo. Our article on whether AI assistants favor big brands covers the limits of these tests.

Documented by the platform. OpenAI says ChatGPT considers “structured metadata from first-party and third-party providers (e.g., price, product description) and other third-party content,” and builds review summaries from reviews on public websites. It also says a budget in the question shifts the focus to price.

Observed in our studies. Answers move. In our study of repeated questions, one ChatGPT answer showed 57.8% of the brands its five answers to the same question named between them, and only 25.2% of those brands appeared in all five. Retail questions were among the least stable, with a mean overlap of 0.458 between runs. Video also plays a part in retail answers: in our YouTube study, Google’s AI Overviews cited a YouTube video on 45.0% of retail searches, even though they cited only 15.0% of the videos shown on page one.

Our inference for sportswear. The trust signals shoppers use are concrete: fabric weight and opacity, sweat handling, pockets, support level, inseam and sizing range, wash durability, and real wear-testing by runners, lifters or players. When those details are written plainly and repeated consistently by retailers, reviewers and creators, an assistant can match them to a question. Hype words cannot be matched to anything; see our guide to product content AI shopping assistants prefer.

What does a sportswear brand lose when AI names a rival?

The highest-intent shoppers, at the moments when they are ready to buy several items at once.

No one has published how much sportswear revenue AI answers redirect, and we do not estimate it. The evidence supports the direction. Sports moments create bursts of specific questions. Google says it answers those with tailored recommendations and side-by-side comparisons. AI-driven retail visits were worth more per visit than other visits in Adobe’s data. And the leaders’ own digital channels are under pressure in some cases, with Nike’s brand digital sales down 12%. A brand absent from the answer for “best anti-chafe running shorts” loses a shopper who was about to buy, and possibly the rest of the kit that shopper buys alongside.

How does generative engine optimization work for sportswear brands?

By making each product’s performance facts specific and consistent, and confirmed by independent sources across every AI surface.

  • Product pages built around activity questions. State the activity, conditions, fabric, support level, pockets, fit, inseam and size range in plain words. Lululemon’s plan to update its product pages after its homepage and category pages shows where conversion pressure is landing.
  • Feeds and retailer listings that agree. Keep price, stock, sizes and colors accurate in Google Merchant Center, Shopify Catalog or OpenAI’s merchant program, and make sure wholesale partners describe each product the way you do.
  • Agentic readiness. Decide which AI checkouts you want to support directly and which you leave to retailers, and check that product details survive the trip.
  • Independent proof. Wear tests by running, lifting and team-sport publications, honest reviews, and creator videos that say the important details out loud. Outdoor brands lean on expert reviews in the same way; see how outdoor brands get shortlisted.
  • Event calendars. Publish clear guidance ahead of marathon season, the start of school sports, and major tournaments, when questions spike.
  • Measurement that respects variation. Track your products for the same prompts across ChatGPT, Google AI Mode, AI Overviews, Gemini and Perplexity, repeated over time. Our article on how many prompts to track explains why one check misleads.

None of this guarantees a recommendation; assistants decide their own answers. What a brand controls is whether the facts that win a comparison are available, consistent and believable.

Which sportswear questions does the data leave open?

How much sportswear specifically moves through AI answers, and which ending, brand site, retailer or chat, wins most often.

The traffic and conversion figures here are for all US retail. We found no public sportswear-only figure for AI-referred revenue, no published share of sportswear purchases completed inside assistants, and no independent study of how assistants rank activewear. The brand results cited are company reports, and we do not attribute them to AI. The controlled study used invented brands and typed-in product data, which is cleaner than real shopping. So these figures point the way for a sportswear brand; they do not prove a return.

Where should a sportswear brand start?

With an audit of the activity and attribute prompts tied to your best-selling product lines.

List the prompts your buyers ask before a race, a season or a tournament, then check how ChatGPT, Google’s AI features, Gemini and Perplexity answer them, several times each. Note which products are named, whether their details are right, and whether the answer sends shoppers to you, a retailer or a checkout in the chat. If you would rather hand this off, speak with our team: we will map where your products appear for high-intent prompts, where rivals are named instead, and the work most likely to turn those answers into sales. That work, from activity-led product pages to retailer feeds that agree and event-season guidance, is described on our generative engine optimization service page.

Frequently asked questions

Can AI assistants complete a sportswear purchase without a website visit?

Yes, in some cases. OpenAI’s help page describes Instant Checkout for some eligible merchants, though FashionUnited reports OpenAI scaled it back in March 2026. Google offers an agentic “buy for me” checkout, and JD Sports lets US shoppers buy through AI platforms.

Should a sportswear brand sell directly inside AI assistants or leave it to retailers?

It depends on margins, operations and retailer relationships. Either way, your product data must be accurate wherever the assistant reads it, because it may list several merchants for the same item.

Do big sportswear brands always win AI recommendations?

No. In a controlled test, a famous brand won when products were identical, but a small, clearly stated advantage let an unknown brand win most of the time.

Do creator videos matter for AI answers about sportswear?

They can. Our study found Google’s AI Overviews cited a YouTube video on 45.0% of retail searches, and the text shown for cited videos came from what was said in them.

How often should we check our AI visibility?

Regularly, and more than once per prompt. Answers change between runs and over hours, so a single check can miss half the brands an assistant names.

Sources

Free strategy call

Some questions are easier to answer about your own business.

Bring the one that matters most. On a free 30-minute call we’ll take a first look at it and send you a short written read afterward.

Ask about my businessHow we work