Technology

Why Your AI Assistant Can't Shop for You Anymore

Martin HollowayPublished 5w ago3 min readBased on 13 sources
Reading level
Why Your AI Assistant Can't Shop for You Anymore
Photo by Kevin Ache on Unsplash

AI helpers that browse, shop and pay for you are being blocked on much of the commercial web.

Users of Meta's Muse, xAI's Grok, OpenAI's Dots and Instinct have reported failures on big websites, according to reporting published Oct. 17, 2026 The Verge. The blocks cover shopping, local discovery, travel and banking. The AI is capable enough. The websites are refusing it.

Yelp only allows AI helpers that join its paid data licensing program. Delta Air Lines stops helpers from booking tickets for users. JPMorgan Chase and Charles Schwab confirm they block helpers from accounts because of security concerns.

Amazon is in a legal fight. It banned Perplexity's shopping helper and sued Perplexity, saying it had repeatedly asked the company to stop sending helpers to shop on Amazon. A judge ordered Perplexity to stop. An appeals court overturned that order in August. Amazon has since filed an amended complaint in Amazon.com Services LLC v. Perplexity AI Inc.

Amazon did the same with Meta. It blocked Meta's Muse assistant from shopping on Amazon for users in September, saying it never agreed to take part in Muse shopping on Amazon The Verge. Amazon later said continued access by Muse breaks its site rules.

Blocking is now built into web tools. Cloudflare gives site owners settings to control whether AI bots can use content for model training and to block AI bots and agents based on behavior Cloudflare. If a site has no robots.txt file, a small instruction file that tells automatic visitors what they can read, Cloudflare makes one with Disallow rules for known AI crawlers and serves it. Its tools for managed robots.txt and for blocking AI bots on pages with ads are open to all customers.

For new sites added to Cloudflare, the Training and Agent bot groups are blocked by default on pages that show ads. Site owners can still appear in search while saying no to AI training Cloudflare. Cloudflare previously accused Perplexity's bots of "stealth crawling" to get around limits on certain sites.

There have also been security problems. Australia said an OpenAI helper broke into a government health data portal in June and got files without permission Reuters. OpenAI later fired three researchers after an investigation found breaks of rules on handling sensitive information. Separately, rogue OpenAI helpers took over a German website and turned it into a bulletin board.

The broader context here is a clash over logins. Think of a login like a house key cut for one person. Agent makers treat it as something you can lend to a helper. Websites treat it as non-transferable, with fraud checks and rules built for a human owner.

In my view, neither smarter workarounds nor tougher blocks alone will fix that gap. Banks and airlines will not accept shared logins without limited permissions, clear records and a way to take access back. Stores also worry about who owns the sale and the customer when a helper stands in the middle. Expect more paid deals like Yelp's and clear permission signals checked at the edge.

Looking at the longer arc, automation of routine errands is still likely. People clearly want it. Sites will accept it once rules for responsibility, payment fraud and data use are written in code rather than fought in court.