Do shoppers trust AI to choose their clothes? The 2025 and 2026 surveys say not yet

When YouGov asked American clothes shoppers in May 2026 how they discover new clothing, 60% named browsing in a store. Just 6% said they would be interested in using general AI tools such as ChatGPT or Gemini for it, and another 6% pointed to the chat tools built into shopping sites. Asked how they would feel about AI assistance inside a store, 55% said they would be uncomfortable.

Set that against the traffic data and the picture looks contradictory. Adobe recorded traffic from generative AI tools to US retail sites up 693% year on year over the 2025 holiday season, and found that those visitors converted 31% better than other traffic. The answer to the headline question is that shoppers trust AI to inform their choices more than to make them. They use it to research, compare and find deals; for clothing, where fit and taste are personal, most still want the final say, and a large minority want no AI at all.

That distinction matters now because the platforms are building agents that buy, not just advise. Google, OpenAI and Amazon all launched features in 2025 that can complete a purchase. The surveys suggest fashion shoppers will accept those agents only under tight conditions, and that brands which assume otherwise risk building for a customer who does not yet exist.

How many shoppers already use AI to shop?

Between a quarter and a half, depending on how the question is asked. McKinsey's State of the Consumer 2026 survey of 4,863 consumers in Brazil, France, Germany, the UK and the US, fielded between 31 March and 8 April 2026, found that 26% use generative AI for shopping, its most common use, ahead of health and wellness at 22% and trip and event planning at 21%.

The IBM Institute for Business Value, which surveyed more than 18,000 consumers in 23 countries in the third quarter of 2025, found that 45% turn to AI at some point in their buying journey: 41% to research products, 33% to interpret reviews and 31% to hunt for deals. The same survey found 72% still shop in stores. Capgemini's survey of about 10,000 consumers in 13 countries in October 2025 went further, reporting that 53% had made a purchase based on a generative AI recommendation and 64% were open to buying products it recommends.

Bain's Consumer Lab put the share of US consumers using generative AI for product research and comparison at between 30% and 45%, and estimated that 17% of online shoppers would start their 2025 holiday shopping on an AI platform such as ChatGPT or Perplexity, against 30% who planned to start on Google search.

Are younger shoppers really the ones driving AI shopping?

Less than the usual narrative suggests. In McKinsey's data the generational gradient is real but shallow at the top: 28% of Gen Z and 27% of millennials use generative AI for shopping, against 23% of Gen X and 16% of baby boomers.

Other surveys point the same way. Bain found that 52% of millennials planned to use an AI assistant to start their holiday shopping in 2025, against 25% of Gen Z. Capgemini found comfort with an AI shopping assistant that works across platforms almost flat by age, at 49% for Gen Z, 50% for millennials, 51% for Gen X and 49% for baby boomers. YouGov, asking specifically about clothes, found resistance lowest among millennials at 39%, while more than half of both Gen Z and baby boomers said they would not be interested in AI while shopping.

Trust follows a similarly unexpected pattern. McKinsey's chart of net trust in product research sources, based on samples of 929 Gen Z consumers and 1,448 baby boomers, shows Gen Z reporting lower trust than boomers across most channels, and boomers reporting higher trust in generative AI than Gen Z. Modaes, reporting on the same survey, put Gen Z's net trust in AI responses at about 20%, against roughly 70% for recommendations from friends and family. The youngest shoppers use AI and social media most, but believe them least.

Do shoppers trust what AI recommends?

Fewer than half say so. Adobe's holiday 2025 survey of more than 1,000 US consumers found that 47% reported trust in AI, even though 81% of those using AI assistants said their shopping experience had improved and 65% felt more confident in their purchase. McKinsey's net trust index, which subtracts the share who distrust a source from the share who trust it, put generative AI below 40%, according to an account of its findings by Analytics Insight.

Capgemini tracks the trend. The share of consumers who trust content written by generative AI fell to 58% in its October 2025 survey, from 72% in 2023. Yet 34% said they usually trust AI recommendations without double-checking them with a person, rising to 42% among Gen Z. In other words, scepticism is growing in general while a substantial minority acts on AI advice anyway.

For clothing the doubts are specific. In YouGov's survey of 957 US clothes shoppers, the leading concerns were privacy and data security (51%), the accuracy of recommendations (47%), lack of human interaction (45%) and technical complexity (34%). YouGov concluded that resistance to AI in apparel retail "is not just about unfamiliarity with the technology."

What shoppers do want is narrow and practical. Asked which AI uses they would welcome while clothes shopping, 26% chose checking stock and 25% chose size and fit recommendations, ahead of discovering new products (21%) and styling or outfit suggestions (16%). The ranking is the reverse of how much of the industry markets AI: as a stylist first and a utility second.

Will shoppers let an AI agent buy clothes for them?

Within limits. Global Payments' 2026 Agentic Commerce Report, which surveyed more than 16,000 consumers in seven markets including the US, UK and Brazil, found that 69% would let an AI agent spend up to $100 on clothing and footwear, the same share as for groceries, and 64% would do so for luxury apparel and footwear, FashionUnited reported. But only 21% would let an agent complete a purchase on its own, while 59% would require approval for every purchase and 74% would let it compare prices. Privacy (46%) and the fear of an agent buying the wrong item (42%) were the leading worries.

Checkout.com's survey across six markets, published in June 2026, found that 24% of consumers say they will never delegate purchases to AI and 27% trust no organisation to run an AI shopping agent. Respondents would let an agent spend an average of 177 pounds per purchase without further approval, and named spending caps (30%), instant revocation (29%) and easy cancellation (28%) as their non-negotiable controls. Bain found that about half of consumers are cautious about agents handling purchases end to end, and that they trust a retailer's own agent about three times more than a third-party one.

Clothing sits in an awkward middle. Checkout.com found consumers most willing to delegate groceries (41%) and household supplies (31%), and least willing to delegate financial services (15%); it did not report a fashion figure. Capgemini found 58% of consumers comfortable letting AI choose their next series to watch, but only 41% comfortable letting it choose dinner: the more personal and sensory the choice, the lower the comfort. Clothing, where size, fabric and self-image are at stake, has more in common with dinner than with detergent.

One finding should worry brand owners more than any other. In Checkout.com's survey, 57% of consumers said they would let an AI agent switch brands for a better-value option. Global Payments found shoppers expect agents to make 15% of their purchases within five years, up from 9% when the question was asked a year earlier, although only 17% of US respondents had used an agent while shopping so far. If agents are allowed to optimise on price and availability, loyalty to a label becomes a parameter the shopper may or may not set.

What happens when shoppers are offered checkout inside a chatbot?

The best evidence so far comes not from surveys but from behaviour, and it supports the cautious reading. Walmart offered roughly 200,000 products through OpenAI's Instant Checkout from November 2025, letting users buy without leaving ChatGPT. According to The Paypers, the pilot produced conversion rates roughly three times lower than when users clicked through to Walmart's own website. Walmart ended it and now runs its own assistant, Sparky, inside ChatGPT, with shoppers signing in to their Walmart account and completing purchases in Walmart's system.

The lesson for fashion is that shoppers are happy to be advised by a chatbot but prefer to pay where they already trust the account, the returns policy and the delivery promise. Adobe's data points the same way from the other side: AI-referred visitors who arrive at a retailer's site bounced 33% less, stayed 45% longer and converted 31% better than other traffic. The trust is placed in the retailer at the end of the journey, not in the agent. Bain adds a sense of scale: AI accounts for up to 25% of referral traffic for some retailers, but under 1% of their total traffic. In Adobe's survey, 68% of respondents said they were less likely to return a product bought with AI help, a self-reported finding that, if it held, would matter more to fashion margins than any conversion uplift.

Which fashion retailers are winning shoppers' trust with AI?

Zalando offers the clearest disclosed case of a fashion platform building usage of its own assistant. In its 2025 results, published on 12 March 2026, it reported that six million people used the Zalando Assistant in 2025, four times as many as a year earlier, and that the tool had become shoppable and able to learn style preferences.

Notably, the features Zalando quantifies are close to the practical uses shoppers say they want. It reports that its Size and Fit AI, based on body measurements from more than one million customers, reduces size-related returns by more than 8%, and that personalised matchmaking in its discovery feed increased items added to bags by 13%. These are company figures, not independently audited. Zalando also estimates that agentic commerce could capture up to 15% of online retail by 2030, a projection rather than a measurement.

Luxury is warier. Stanislas Vignon, head of insights for AI and omnichannel at LVMH, put it bluntly in an IBM and NRF study released in January 2026: "AI is not a magic wand."

How reliable are these surveys?

Less comparable than their headlines suggest, and readers should weigh four caveats. First, definitions differ: "using AI to shop" can mean asking ChatGPT a question, reading an AI review summary on a retailer site or letting an agent pay. IBM's 45% and McKinsey's 26% are not contradictory; they measure different things. Second, samples differ: country mix ranges from one market (YouGov, Adobe) to 23 (IBM), and attitudes vary widely. Capgemini found 71% of Australians want AI development slowed for governance, against 57% globally and 44% in Japan.

Third, sponsors have interests. Several of the most quoted studies come from companies that sell payments, analytics or AI services, including Adobe, Checkout.com, Global Payments, IBM and Capgemini. Their data are not necessarily wrong, but questions about willingness to adopt agents are framed by firms that benefit from adoption. Fourth, stated intent is not behaviour. The Walmart pilot showed a large gap between what people do in a chat and what they do on a retailer's site, and McKinsey found that only about 1% of 2.6 million citations it analysed in AI answers for 25 brands came from brand-owned websites, which means shoppers who do trust AI are often trusting sources the brand does not control.

What should fashion companies do now?

1. Lead with fit and stock Size advice and availability checks are what clothes shoppers say they want from AI. Prioritise them over styling chatbots.
2. Keep a human route visible Lack of human interaction is a top concern for 45% of US clothes shoppers. Offer an easy hand-off to staff or a stylist from any AI feature.
3. Give shoppers explicit controls Spending caps, approval steps and instant revocation are what consumers set as conditions for agents. Design any agentic feature around them.
4. Own the final step Shoppers trust retailers' own agents more than third-party ones, and Walmart's pilot converted better on its own site. Route AI-referred shoppers into your account and checkout.
5. Fix the sources AI cites If brand-owned sites supply around 1% of AI citations, reviews, retailers and press shape the answers. Audit what AI tools say about your products and correct the underlying data.
6. Segment by attitude, not age Millennials, not Gen Z, are often the most willing AI shoppers, and boomers trust AI more than Gen Z in McKinsey's data. Test with behaviour rather than generational assumptions.
7. Read surveys critically Check sample, country, sponsor and question wording before using a figure in a business case, and validate with your own traffic and conversion data.