• AI visibility

Guide · AI search

How can a footwear brand get its shoes found when shoppers ask AI what to buy?

By making each shoe model easy for AI assistants to match to a shopper’s fit, use and budget question, and easy to verify on independent sites and retailer listings. Footwear is bought model by model, and comfort and fit come first, so the work happens at the level of the Clifton or the Cloudmonster, not the logo. This guide covers who decides, where AI now sits in the shoe-buying journey, and what a brand can and cannot control.

By Underneath Team Guide Updated

The short version

  1. The prize is large and moving toward performance: US footwear sales held at $90 billion in 2025, with running up 9% and walking up 9% in dollars, according to Circana (opens in a new tab); in the first half of 2026, running shoes rose 13% in value and volume (World Footwear (opens in a new tab)).
  2. Shoppers decide on comfort and fit first, then quality, then price, in a Footwear Insight survey (opens in a new tab) of 307 active consumers. Those are exactly the questions people now type into AI assistants in full sentences.
  3. Awareness is still the constraint even for fast-growing brands: On (opens in a new tab) reports global brand awareness of 30% while its own channels reached 45.7% of second-quarter sales.
  4. Much of the sale still happens at retailers: Deckers’ wholesale sales were $3.21 billion against $2.26 billion direct to consumers in fiscal 2026 (FashionUnited (opens in a new tab)). An AI answer can send the shopper to either.
  5. Google added shoes to its AI “try it on” tool in October 2025 (Google (opens in a new tab)), and OpenAI documents that ChatGPT ranks the merchants for a product partly on whether they are “the maker or primary seller.”

Who decides which shoes get bought, and what is one customer worth?

The shopper decides, model by model, and a good fit often turns into years of repeat pairs.

For a footwear brand, the buyer is not a procurement committee. It is a runner replacing worn-out trainers, a nurse who needs to survive 12-hour shifts, a parent buying school shoes, or someone chasing a new lifestyle sneaker. What they weigh is consistent. In the Footwear Insight survey, run on the MESH01 panel of active and outdoor consumers, comfort and fit ranked first, quality and durability second and price and value third. The same shoppers said they were most likely to buy running shoes (53%), casual or comfort shoes (47%), walking shoes (43%) and boots (43%) that fall and winter.

Price pressure is real. 66% of those shoppers had seen price increases on footwear they wanted, and 50% said they would wait for sales if prices kept rising. Circana noted that nearly half of consumers have delayed purchases or chosen cheaper alternatives because of price increases. A shopper who is stretching a budget researches harder before committing. Furniture shoppers spending hundreds to thousands of dollars research across several visits and channels, as our guide on furniture brands winning high-ticket orders explains.

What a customer is worth depends less on one order than on the model they settle into. Runners and people who stand all day tend to rebuy a shoe that works. That habit is why the growth brands report the way they do: Deckers’ HOKA brand grew 15.9% to $2.59 billion in the fiscal year to March 2026, and On’s shoe sales reached CHF 781.6 million in a single quarter. Our inference is that the first match between a shopper and a model is the valuable moment, because it can set the next several purchases.

Executives should keep the channel split in mind. At Deckers, wholesale sales grew 12.3% and direct sales 6.3% last fiscal year. At On, direct sales grew 26.0% in the second quarter, and the company credits its rising direct share, along with full-price discipline, for a 65.4% gross margin. So the same AI answer can feed a higher-margin sale on your site or a sale through a retail partner, depending on where the shopper clicks. How a retailer competes for that same click, with accurate prices and stock, is shown in our guide on how beauty retailers turn AI picks into sales.

Where do AI assistants already sit in the shoe-buying journey?

Between the need and the shortlist, where shoppers used to ask a store specialist or read reviews.

The cross-retail evidence is strong. Adobe’s data (opens in a new tab), reported by TechCrunch, shows AI traffic to US retail sites rose 393% in the first quarter of 2026, and that this traffic converted 42% better than other traffic in March 2026. Adobe’s survey found 39% of people had used AI for online shopping. These are all-retail figures, not footwear figures, but they describe the same shoppers.

Google and OpenAI have both built shopping features that suit shoes. Google says its Shopping Graph now has more than 50 billion product listings, with more than 2 billion refreshed every hour, and its AI Mode shopping announcement (opens in a new tab) opens with a shoe question: “How do I choose a pair of hiking boots?” In October 2025 Google extended its try-on tool to shoes so shoppers can see “those heels or sneakers” on a photo of themselves (TechCrunch (opens in a new tab)). OpenAI’s help page (opens in a new tab) describes product results with images, prices, review summaries and links, plus a try-on button for clothes and accessories that, it warns, does “not guarantee fit or size.” Our guide on how fashion brands get matched to a look covers these visual and try-on tools for clothing.

Our own data shows how often Google already answers before the click. In our study of when Google shows AI Overviews, 68.0% of the 100 retail and ecommerce keywords showed one, and 95.0% of the question-form retail keywords did. Footwear questions are often questions: which shoe, for what, in which size.

What do shoe shoppers ask AI assistants?

Long, specific questions about fit, use, comparisons, sizing and price, usually naming a problem rather than a product.

We wrote the examples below to illustrate the kinds of questions footwear shoppers ask; they are not observed prompts:

  • Use case: “Best cushioned running shoe for a heavier runner training for a first marathon.”
  • Fit: “Which running shoes have a wide toe box but still feel light?”
  • Comparison: “Hoka Clifton vs Brooks Ghost for someone who walks 10 miles a day at work.”
  • Sizing: “Does this model run small compared with my usual size in Nike?”
  • Work and lifestyle: “Comfortable black shoes for a restaurant server that pass a dress code.”
  • Weather: “Waterproof sneakers that don’t look like hiking boots.”
  • Price: “Best walking shoes under $100 that last more than a year.”

Research on AI shopping suggests these questions are long for a reason. Bagga and colleagues (opens in a new tab) built a set of 13,747 shopping requests modeled on how people describe needs on Reddit; they averaged about 59 words, against three or so words for typical search keywords. Shoppers state budgets, past experiences and must-have features. A shoe page that only says “responsive cushioning” gives an assistant little to match against “my knees hurt after long shifts on concrete.”

Some of these questions touch on foot pain or injury. Brands should describe construction and intended use plainly and leave diagnosis to clinicians. That is good practice in any channel, and claims that sound medical invite more scrutiny, not less.

How does an AI answer turn into a pair sold?

Through a model named on a shortlist, a size decision, and a click to your site or a retailer.

The path in footwear has four steps, and each one can be lost.

Named. The assistant suggests a few models for the stated need. If your shoe is not among them, nothing after this point happens. On’s 30% global brand awareness, reported at a time of strong growth, is a reminder that most shoppers anywhere still do not know most brands; an answer that names a model is a way to be introduced. Luxury brands see the same pattern, since most of their shoppers’ AI questions name no brand at all, as how luxury brands reach high spenders shows.

Described correctly. The shopper reads how the assistant characterizes the shoe: cushioned or firm, narrow or roomy, true to size or not. A wrong description loses the sale, or worse, wins a sale that comes back as a return.

Sized. Fit is the top purchase factor, so the shopper needs a confident size. Try-on tools help with looks, but both Google and OpenAI frame them as visual. Clear size guidance on your own pages and retailer pages is what the answer can repeat. Our guide for apparel brands on fit and size facts covers the same problem for clothing.

Bought. OpenAI documents 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 reasonable expectation is that a brand with accurate stock and price data on its own store can capture more of these clicks, while a brand whose retailers have better data will see the sale go to them. Either is revenue, at different margins.

Shopify reports results along this path for its merchants. Its 2026 holiday report (opens in a new tab) says 65% of shoppers plan to use AI for at least one shopping task this season, and notes that only one-third trusted an agent to buy for them. In other words, AI shapes the choice, and the shopper still completes the purchase.

What decides whether a shoe model gets named?

The platforms document product data and merchant factors; studies point to verifiable facts; the rest is our inference.

Documented by the platforms. OpenAI says ChatGPT considers “structured metadata from first-party and third-party providers (e.g., price, product description) and other third-party content,” and that review summaries are built from reviews on public websites. For Shopify merchants, OpenAI says product data already reaches ChatGPT through Shopify Catalog. Google says AI Mode runs several searches at once, a “query fan-out,” to work out what makes a product good for a stated situation, then suggests options against those criteria.

Observed in studies. In a controlled test summarized in our article on what drives AI product picks, product facts such as rating, price and reviews mattered far more than brand name when the assistant could see them. Bagga and colleagues found that rewrites that kept facts, named concrete attributes and answered likely buyer questions moved products up, while advertising-style copy did not. For a shoe page, that means fit and construction facts over taglines, as our guide to the product content AI shopping assistants prefer explains.

Our inference for footwear. The trust factors shoe shoppers use are specific: width options, heel-to-toe drop, weight, cushioning, outsole grip, water resistance, how a model sizes against other brands, how long it lasts, and what changed between versions. When those facts match across your product page, your retailers’ listings and independent reviews, an assistant has consistent material to work with. When version 9 and version 10 of a shoe are described differently on different sites, a reasonable expectation is that answers will blur them. New models face an added delay; our article on why ChatGPT misses new products explains why launches can be invisible for a while.

What does a footwear brand lose when AI leaves its shoes out?

The first introduction to a shopper who might have bought the same model for years.

We have no published measure of lost footwear sales from AI answers, and we will not invent one. What the evidence supports is the shape of the risk. Running and walking are where US footwear growth is, according to Circana. Those are the categories where shoppers ask detailed fit and use questions. AI-referred shoppers converted better than other traffic in Adobe’s retail data. A brand left off the shortlist for “best shoe for standing all day” therefore loses the shoppers most ready to buy, in the categories that are growing.

There is a second, quieter cost: a shoe named with the wrong fit or features can turn into returns and poor reviews, which then feed the next answer. Because answers change from run to run, a single check tells you little; our article on why AI answers about your brand change covers how to measure it properly.

How does generative engine optimization work for a footwear brand?

By making each model’s fit, use and price facts clear, consistent and confirmed by independent sources wherever AI looks.

The work is practical and model by model:

  • Model pages that answer fit questions. Width options, drop, stack height, weight, sizing advice against common reference brands, and the use each shoe is built for, written in plain words a shopper would use.
  • Clean product data everywhere it flows. Feeds to Google Merchant Center, Shopify Catalog or OpenAI’s merchant program, with live price, stock and every size and width, so assistants can list you as the maker with accurate availability.
  • Retailer listings that match. Many sales still happen at wholesale partners. Their pages should describe each model the same way you do.
  • Independent coverage. Lab-style reviews, specialty running and workwear publications, and honest customer reviews give an assistant something to verify against. Shopify quotes Stanley 1913’s commerce lead saying that this kind of optimization means positioning a brand’s story “in other authoritative sources.”
  • Version clarity. Name each new version, say what changed, and keep older versions’ pages accurate.
  • Visibility tracking across assistants. Measure which models are named, for which use cases, in ChatGPT, Google AI Mode, AI Overviews, Gemini and Perplexity, repeated over time.

None of this can guarantee a recommendation. No one controls what an assistant says. What a brand controls is whether the facts it needs are available, consistent and believable.

What don’t we know yet about AI and shoe sales?

No one has published how often shoe shoppers use AI, or how many pairs it sells.

The AI shopping figures here come from all of retail, not footwear alone. We found no public footwear-specific measure of AI-referred sales, return rates for AI-referred orders, or how often assistants name specific shoe models. Platform documentation describes product results in general, not how shoes are ranked. Try-on tools are new, and no one has published whether they change size choice or returns. Brand results such as Shopify’s merchant stories are self-reported. Read these numbers as pointing the way for footwear, not as measured shoe results.

Where should a footwear brand start?

With an audit of how AI assistants describe your best-selling models for the fit and use questions buyers ask.

Pick the five or ten models that carry your revenue, list the use-case, fit and comparison questions their buyers ask, and check what ChatGPT, Google’s AI features, Gemini and Perplexity say, run several times. Look for three things: whether each model is named, whether its fit and features are described correctly, and which store the answer sends shoppers to. If you would like a hand, send us your top models and we will map where your models appear, where they are missing or misdescribed, and the work most likely to put them on more shortlists that end in a sale. Our generative engine optimization service page explains how that model-by-model work runs, from fit-focused model pages and clean feeds to matching retailer listings and clear version names.

Frequently asked questions

Do AI assistants recommend specific shoe models or just brands?

Shopping features in ChatGPT and Google show individual products with prices, images and links, so the unit is usually the model. That is why model pages, sizes and versions need accurate, consistent facts.

Can AI try-on tools tell shoppers which size to buy?

No. Google’s try-on shows how shoes look on a photo, and OpenAI warns that its try-on images “do not guarantee fit or size.” Size advice still has to come from your pages, retailer listings and reviews.

Will AI shopping send sales to my retailers instead of my own store?

Sometimes. OpenAI says ChatGPT ranks merchants on availability, price, quality and whether the seller is the maker. Accurate stock and price data on your own store helps, and retailer sales still count as revenue.

Should footwear brands make health claims to match pain-related questions?

No. Describe construction and intended use plainly, and leave medical advice to clinicians. Claims that sound medical attract scrutiny from regulators, retailers and shoppers alike.

How long before changes show up in AI answers?

It varies by assistant and by source. Product data can update quickly; independent reviews and articles take longer to appear and be picked up. Track the same questions repeatedly rather than checking once.

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