This week, we take a look at how some fashion retailers and consumers are responding to the rise of AI shopping agents. While the technology is growing in adoption, both retailers and consumers have concerns about security, fraud, reliability and the loss of advertising revenue.
As AI-powered shopping agents increasingly become a part of modern e-commerce, not everyone is happy about their growth.
Amazon is in the midst of a legal battle against the AI company Perplexity over Perplexity’s use of AI agents that can browse and scrape elements from Amazon’s website. Earlier this month, Amazon won a court order blocking Perplexity from accessing Amazon, citing both privacy concerns for users and adverse effects to Amazon’s advertising business.
Elsewhere, Target issued a warning this week to consumers about its upcoming partnership with Google’s Gemini, specifically noting that any product purchased by Gemini on the user’s behalf would be “considered transactions authorized by you.” In other words, customers will still be on the hook for any transactions, even if Gemini erred in its decisions.
Both of these incidents reflect the two sides of concerns about AI shopping agents. For retailers, having another company’s AI agents digging around in your website presents potentially expensive logistical issues. For consumers, the experience of using a shopping agent is still fraught with the potential for costly mistakes and incorrect orders.
Upasna Singh, the founder and CEO of the online fashion boutique U-Glam, told Glossy that she has experienced both sides herself. As a customer, she said shopping for fashion is “personal and intuitive” and that AI agents can miss out on the emotional connection a customer may have to a piece before they buy it.
“As an e-commerce owner, the concerns are much more immediate,” Singh said. “I deal with bot traffic regularly, whether it’s scraping designs I’ve spent months perfecting or attempting to access sensitive areas like payments and customer data. These are ongoing realities, not theoretical risks.”
Retailers rely on the data they get from customers shopping their online stores. That data, including what items catch people’s eye, helps guide decision-making. But AI agents complicate that picture and require retailers to spend more time filtering through their traffic to determine what came from a real person.
Amazon’s moves make sense when you understand what’s actually at risk,” said Sayali Patil, AI infrastructure product leader at Cisco. “It’s not just the $56 billion ad business, it’s data ownership. When a customer shops through an external agent, retailers lose the behavioral data trail: what was searched, what was viewed, how long they lingered. That’s the real asset being protected, and I’d expect more retailers to follow suit as they realize agents don’t just disintermediate the transaction, they disintermediate the relationship.”
On the consumer side, many of the headaches that come with shopping agents, including incorrect product information and prices, come down to the way the agents access a website. According to Burkan Bur, managing director and head of SEO at the digital agency The Ad Firm, these agents are accessing cached webpages rather than the up-to-date version a consumer would see in a web browser.
“Shopping agents don’t refresh product data, and they access cached versions that are sometimes hours old, sometimes scraped from a page that has changed twice since the last time it was crawled,” Bur said. “In some fast-moving product catalogs, the price can change 40-60 times a day. When the consumer takes action on that and gets the wrong price or an out-of-stock product, no fingers are pointed at the agent, and the consumer complains to the retailer. Amazon’s lawsuit is a natural outgrowth of someone using your product listings to make money from billions of product impressions.”
With the release of more AI agents that can spend money and make purchases, like OpenClaw , there is more need for security and guardrails. But those guardrails are slower to catch up to the adoption of AI shopping agents. At the beginning of March, Mastercard’s vp of government affairs and policy and its evp of cybersecurity released an open letter calling for international standards for security around AI agents to protect against fraud.
Until clearer security standards are in place, AI agents will likely remain just as much of a risky prospect as an exciting opportunity.
“Ultimately, the biggest obstacle to adoption is trust,” Singh said. “Consumers need confidence that AI will act in their best interest, and businesses need assurance that these systems won’t compromise security, ownership, or compliance.”
AI agents, by the numbers
- According to a survey of 2,500 shoppers who do not use AI to shop by the consultancy Simon-Kucher, their No. 1 reason was a lack of personalization in gift-giving. The second most common reason was a lack of trust in AI’s capabilities. This data was provided to Glossy by Simon-Kucher.
- 70% of fashion-related questions asked to AI shopping agents are related to fit and sizing, according to the Boston-based AI agent company True Fit.
- 53% of customers who use generative AI for search also use it to help them shop, according to the 2026 BoF-McKinsey State of Fashion report.
- Ipsos released a report earlier this month with many interesting data points about consumer sentiment toward AI shopping agents. For one, the majority of consumers want to approve every purchase every time and are uncomfortable with the idea of an agent making a purchase with no approval.
Other news to know
- LVMH’s share price fell by 28% in the first quarter of the year, marking the company’s biggest drop to date. The company largely blamed the war in Iran for its slump.
- U.S. Customs and Border Protection is progressing its tariff refund system. A filing on Tuesday said refunds would likely take up to 45 days after the application is submitted to be fulfilled.
- Allbirds is being acquired by American Exchange Group for $39 million. The company was once valued at $4 billion.