Only 7% of Consumers Trust AI to Shop Autonomously, Putting Retail Trust at Stake

Only 7% of consumers in the fashion and sportswear sector are willing to let artificial intelligence (AI) assistants complete transactions without their approval. This finding comes from an ACI Worldwide survey of over 3,300 respondents in the United States and the United Kingdom, released last week, revealing a gap between enthusiasm for AI and the readiness to relinquish control over one's own wallet.

This figure is not an anomaly. At least 53% of respondents admitted discomfort with AI shopping on their behalf, while another 14% required manual approval for every purchase made by a digital assistant. This implies that AI adoption in retail currently largely stops at the recommendation stage rather than transaction execution.

Interestingly, consumers still see great value in AI. The most appreciated features are price drop alerts and cross-retailer price comparisons. They trust algorithms to find the best deals but have not yet surrendered the final decision. "I don't think anyone wants to give up that will," said Marie Driscoll, President of the Retail Marketing Society, to CX Dive.

Isabelle Zdatny, Head of Thought Leadership at Qualtrics XM Institute, believes the business opportunity lies not in "AI shopping for me," but rather in "AI helping to filter options" with the final decision remaining in the consumer's hands. Privacy and trust are major obstacles: more than half of global consumers worry AI endangers personal data, and only two out of five people are confident that companies manage their information responsibly.

Driscoll highlights a broader shift in the trust landscape. Referring to the Edelman Trust Barometer, she states that society no longer gives blind trust to brands. "Loyalty is a function of the quality of output delivered by staff, retailers, websites, or AI agents," she asserted. The implication is that customer retention is no longer determined by brand name, but by the consistency of the experience designed by AI itself.

"You don't want AI in a fashion environment just presenting price comparisons. That's too much like an economist. What you need is a better editor. Explain why that choice is superior," said Marie Driscoll, President of the Retail Marketing Society.

In the fashion context, consumers want AI that can curate catalogs, not just a price comparison engine. Driscoll refers to this role as a "better editor"—AI capable of explaining the superiority of materials, the comfort of cuts, the elasticity of fabrics, and even compatibility with items consumers already own. This ability to tell a story and provide reasons is what distinguishes trusted AI from ignored AI.

For the Indonesian market, these findings serve as an important mirror. The rapidly growing e-commerce ecosystem in the country—supported by Shopee, Tokopedia, and TikTok Shop—has long utilized algorithm-based recommendations. However, features like automated purchases or smart subscriptions still face cultural resistance: Indonesian consumers tend to want to see, weigh, and decide for themselves. The biggest opportunity lies not in full AI autonomy, but in assistants that accelerate product research and offer personalized discounts without taking over transactions.

Retailers capable of designing AI as a smart editor—transparent about data, providing reasons behind every recommendation, and offering a manual approval button—have the potential to win loyalty amidst a crisis of trust. Conversely, those who force full automation risk losing increasingly privacy-aware customers. The question now is: will the industry choose a collaborative path that respects consumer control, or get caught in a feature race that only widens the trust gap?