• AI visibility

Guide · AI search

How does a fashion brand get recommended when shoppers ask AI what to wear?

By being easy to match to a look, an occasion and a style: product data written in the words shoppers use, images that AI try-on and visual search can work with, and the same details on every retailer that carries the brand. Google, ChatGPT, Pinterest and retailers such as Zalando now offer AI styling and try-on tools. Shopper use is real but uneven, so fashion brands should build for it now without expecting it to replace social and stores yet.

By Underneath Team Guide Updated

The short version

  1. Fashion is one of the biggest online categories. US shoppers spent $49.0 billion on apparel (opens in a new tab) online over the 2025 holidays, up 7.4%, Adobe found.
  2. The AI tools are built for style. Google says (opens in a new tab) shoppers can virtually try “billions of apparel listings” on themselves, and OpenAI documents (opens in a new tab) a “Try on” button for clothing in ChatGPT.
  3. Retailers’ own assistants are scaling. Zalando (opens in a new tab) says close to 10 million customers asked its Assistant for advice in early 2026, up from 6 million in all of 2025.
  4. US shoppers are cautious. In a YouGov poll (opens in a new tab), just 6% of clothes shoppers said they would use ChatGPT or Gemini to discover clothing, against 60% who prefer browsing in stores.
  5. Inconsistent product details are common. In a UK survey (opens in a new tab) of 2,000 consumers, only 14% said fashion product details always match across social platforms, marketplaces and retailer sites.

Who shops for fashion online, and how do they find what to buy?

Fashion shoppers start with inspiration, not a product name, and decide on look, occasion and price.

A fashion purchase rarely begins with a spec. It begins with a wedding invitation, a trip, a trend seen on TikTok or a feeling about a style. That is why fashion discovery has lived on social platforms, Pinterest boards, editorial edits and store windows, and why search for clothing is so often visual. In the Connected Consumer 2026 report (opens in a new tab) by Athos Commerce and Drapers, 62% of UK consumers browse for fashion online at least weekly but only 38% buy that often. Most of the journey is looking.

The business stakes are high and the climate is hard. Over the 2025 holidays, apparel discounts peaked at 25.1% off list price in Adobe’s data (opens in a new tab). The State of Fashion 2026 report (opens in a new tab) by McKinsey and The Business of Fashion, as summarized by the CFDA, found nearly half of executives expect industry conditions to worsen in 2026, a share up 8 percentage points on the year before, and executives named artificial intelligence as the industry’s biggest opportunity.

Fashion brands also sell through many doors: their own site, department stores, multi-brand platforms such as Zalando, and marketplaces. Each has its own product pages, and each can be the page an AI tool reads.

Where do AI shopping tools fit into fashion discovery today?

They sit at the inspiration and narrowing stage, increasingly with visual and try-on features.

Four kinds of tools now help shoppers find fashion:

  • Google AI Mode. Google says its AI Mode shopping experience draws on a Shopping Graph of more than 50 billion product listings, more than 2 billion of them refreshed every hour, and shows “a beautiful, browsable panel of images and product listings” for style requests. Its virtual try-on works with the shopper’s own photo for shirts, pants, skirts and dresses.
  • ChatGPT. OpenAI documents a “Try on” button on clothing and accessory listings and says shoppers can upload a photo of clothing and ask how it would look on them. It warns that try-on images “do not guarantee fit or size.”
  • Pinterest. Pinterest reported 640 million monthly users in the second quarter of 2026, and its CEO said its AI is “trained on our unique human curation of style and taste.” Its Pinterest Assistant (opens in a new tab) takes voice, text and image input and returns shoppable results.
  • Retailer assistants. Zalando’s Assistant now covers fashion and beauty in one conversation. Zalando had 62.3 million active customers in the first quarter of 2026.

How many shoppers use these tools depends on who you ask. YouGov’s May 2026 poll of US clothes shoppers found 6% would use general AI tools to discover new clothing or brands, and 16% were interested in styling or outfit suggestions. The UK Connected Consumer survey found 60% of consumers use tools such as ChatGPT, Claude or Gemini at least occasionally while shopping for fashion, and 38% trust fashion recommendations from AI. The two surveys asked different questions in different countries, so the honest reading is that AI fashion discovery is growing from a minority base.

Which questions do fashion shoppers ask AI?

They ask about occasions, looks, trends, brands like the ones they love, and whether a brand is worth it.

The prompts below are illustrative, written to show the kinds of request fashion shoppers make; they are not observed data.

TypeExample prompt (illustrative)
Occasion“What should I wear to a fall wedding as a guest?”
Look or aesthetic“Quiet luxury outfit ideas under $300”
Styling“What tops go with wide-leg jeans?”
Trend“Is the burgundy trend still in this winter?”
Similar brands“Brands like [label] but more affordable”
Quality check“Is [brand] good quality for the price?”
VisualA photo of a coat seen on the street: “Where can I buy this?”

Google’s own example is the second type: a shopper asks AI Mode for “a cute travel bag,” and AI Mode runs several searches at once to work out what makes a bag good for a rainy trip before suggesting options. Visual queries are already common. In the UK survey, 58% of consumers had used visual search to look for fashion items, and 52% went on to buy something afterward; among Gen Z, 73% had used visual search to discover fashion.

How does an AI recommendation turn into a fashion sale?

The tool turns a style request into a shortlist of looks, the shopper tries or compares, then buys.

The path runs from an occasion or style prompt to a visual set of options, through a try-on or comparison, to a product page and a basket. Two features make it different from other categories.

First, the answer is a set of images. Google describes AI Mode showing an image panel that updates as the shopper refines the request; ChatGPT shows product carousels with imagery. We infer that a garment shown on a clear, well-lit image, with the color and silhouette named in the data, has a better chance of being matched to a visual request than one with a vague title. Home decor brands face the same look-first requests, as our guide to decor discovery in AI tools explains.

Second, the sale often lands with a retailer, not the brand. When the shopper asks Zalando’s Assistant, it recommends from Zalando’s catalog. When they ask ChatGPT, OpenAI says the merchant list is ranked partly on “whether they are the maker or primary seller of that item,” alongside price, availability and quality. A fashion brand’s own site can win that click, but only if its stock and price are competitive.

Shoppers still want control of the final step. A Global Payments survey (opens in a new tab) reported by FashionUnited found 69% of respondents would let an AI agent spend up to $100 on clothing and footwear, but 42% worried an agent could buy the wrong item, a risk FashionUnited notes is high in clothing because sizing varies by brand. Our guide for apparel brands on fit and size facts covers the details AI tools need to match clothes to a shopper.

What decides which fashion brands and products AI recommends?

Platforms document their inputs only in part; studies observe a few fashion patterns; the rest is inference.

Documented by the platform. OpenAI says ChatGPT considers structured product data such as price and description from first-party and third-party providers, “other third-party content,” and public reviews, and that product results are not ads. Google says the Shopping Graph holds details like reviews, prices, color options and availability.

Observed in a study. A 2026 measurement of 55,393 trending Google searches found AI Overviews appeared on only 3.5% (opens in a new tab) of Beauty and Fashion queries, the lowest of 19 categories, against 46.1% for Hobbies and Leisure. When they did appear, user-generated platforms, led overall by YouTube, Facebook and Instagram, supplied 28.9% of Beauty and Fashion citations, the highest share of any category. Those were trending queries, not shopping searches, so treat this as a signal that fashion answers lean on social content rather than proof. Our own study of when Google shows an AI Overview looks at the same question for commercial searches.

Our inference. Fashion shoppers describe what they want in style language: “flowy,” “oversized,” “old money,” “wedding guest.” A reasonable expectation is that products whose titles, descriptions and attributes use that vocabulary, alongside color, silhouette, fabric and occasion, are easier for AI tools to match. Creator content and fashion editorial likely carry weight because they supply the style language and the social proof that AI answers draw on. Our guide to what product content AI shopping assistants prefer covers the research on product copy, and does social media help AI search visibility covers which engines cite social platforms.

What does a fashion brand lose if AI tools cannot match it to a look?

It loses the inspiration moment, which in fashion is where most brand choices are made.

No study yet measures lost fashion sales from AI invisibility, so we describe the exposure rather than size it. A shopper who asks for “brands like” a label they already know, or for a look for an event, gets a short set of options. A brand that is not in that set is not considered, and the alternatives often sit on the same retailer pages.

Inconsistent details make the problem worse. When only 14% of UK shoppers say product details always match across channels, AI tools pulling from several retailers may describe the same dress with different colors, prices or fabrics. Seasonal drops add timing risk: AI assistants often answer from older information, which why ChatGPT misses new products explains. A collection that sells for eight weeks can be over before an assistant learns it exists.

There is a brand cost too. A Shopify report (opens in a new tab) quotes Aviator Nation’s ecommerce director on selling $200 hoodies: “There needs to be some emotional connection as to why you will convert and buy this hoodie.” If the story behind a label is not written where AI tools can read it, the answer may reduce the brand to price.

How does GEO work for a fashion brand?

GEO makes your products and brand story easy for AI tools to match, show and trust, without promising placement.

Generative engine optimization (GEO) for fashion covers six areas:

  1. Style and occasion language in product data. Name the silhouette, aesthetic, color, fabric, fit type and occasions in titles, descriptions and feed attributes, using the words shoppers type, not internal style codes.
  2. Images built for visual search and try-on. Clear, consistent on-model and flat images, with descriptive alt text, so visual search and try-on tools can work with the garment.
  3. Consistent details across every door. Make sure wholesale partners, Zalando, department stores and marketplaces carry the same names, colors, prices and fabric details as your own site.
  4. Fashion coverage and creators. Earn editorial edits, stylist features and creator content that describe your pieces in style language. Fashion answers lean on social and editorial sources.
  5. Brand story where AI can read it. Publish plain-text pages on what the brand stands for, who it dresses and how pieces are made, so an answer can explain why a shopper might choose you.
  6. Seasonal timing and measurement. Get new collections into feeds and onto readable pages before launch, then test occasion and style prompts in ChatGPT, Google AI Mode, Gemini, Perplexity and Pinterest, repeating each several times.

What can’t fashion brands learn from today’s AI shopping data?

How many fashion sales AI tools drive, and how they weigh style against price and reviews.

Surveys disagree on how widely fashion shoppers use AI, partly because they ask different questions in different countries. The UK survey comes from a commerce search vendor; YouGov’s poll is independent but asked about interest, not behavior. Zalando’s and Pinterest’s figures are their own. No platform has said how it ranks fashion items for a style request, and the trending-query study did not test shopping searches. Buying inside the chat is also unsettled: FashionUnited reports (opens in a new tab) OpenAI scaled back Instant Checkout in March 2026 after users researched products in ChatGPT but rarely completed purchases there, though OpenAI’s help page still describes it for some eligible merchants. Whether virtual try-on reduces fashion returns has not been shown in published data.

Where should a fashion brand start?

Start with the occasion and style prompts that should lead to your hero pieces and your next drop.

List the occasions, looks and “brands like” comparisons that bring your customers to you, and write them the way a shopper would ask. Run them in ChatGPT, Google AI Mode, Gemini, Perplexity and Pinterest, and on the retailer assistants that carry you. Note whether your pieces appear, which image is shown, which retailer gets the click, and how your brand is described. Then compare your product details across your own site and your three biggest retail partners.

If you would rather have a second pair of eyes on your collections, talk to us about a fashion review. We will show how AI shopping and visual search tools describe your pieces today and which changes are most likely to get them chosen when shoppers ask what to wear. Our generative engine optimization service page walks through the ongoing work, such as writing style and occasion language into feeds, aligning retailer details and timing each seasonal drop.

Frequently asked questions

Do people really use ChatGPT to find clothes?

Some do, but fewer than social media or stores. In YouGov’s US poll, 6% of clothes shoppers would use general AI tools to discover clothing, while a UK survey found 60% use AI tools at least occasionally when shopping for fashion.

Does virtual try-on in Google or ChatGPT help my products get recommended?

Try-on is a shopping feature, not a ranking factor that either company has documented. Clear product images and complete listings make your items usable by it.

Should a fashion brand optimize for Pinterest as well as ChatGPT?

Yes, if your customers use it. Pinterest reports 640 million monthly users and offers an AI assistant for shoppable results, and visual discovery is central to fashion.

How fast do AI assistants pick up a new collection?

It varies. Feeds and readable product pages can reach shopping tools quickly, but assistants that answer from older information may miss new pieces, so launch data early.

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