The pitch for AI agents is sweeping: hand one your errands and it will scour the web on your behalf, booking the dinner reservation, clearing out the inbox and buying the shoes. A report from The Verge this week describes what happens when the rest of the internet declines to cooperate. In recent weeks, users of Meta's Muse, xAI's Grok, OpenAI's Dots and Instinct's agent have run into blocks that stop their agents from completing tasks across major websites, and the companies behind several of those sites say the restrictions are deliberate.
The clearest lines have been drawn around money and travel. Yelp told TechCrunch that it only allows AI agents that pay for its data licensing program. Delta Air Lines is preventing agents from booking tickets on behalf of users. JPMorgan Chase and Charles Schwab have both confirmed that they are blocking AI agents from accessing customer accounts, citing security concerns. Some users say sites that once worked with their agents, including Cars.com and Amtrak, have quietly stopped doing so, although neither company has confirmed a deliberate ban.
Amazon has been fighting this battle the longest. Last year it banned Perplexity's agentic shopping feature and sued the startup, saying it had repeatedly asked Perplexity to stop and that companies buying for customers should operate openly and respect a provider's decision about whether to participate. Perplexity answered that Amazon does not want agents roaming its platform because agents do not look at ads, likening an agent to an extension of the user. A judge initially ordered Perplexity to halt agent shopping on Amazon; an appeals court overturned that ruling in August, and Amazon has since filed an amended complaint. Meta's Muse has hit the same wall, with Amazon saying its "continued access" violates site policies, and Grok, Instinct and Dots agents have reportedly failed on Amazon too.
The stakes go to the heart of what agents are sold as. "Does this impact the perceived value of the Anthropics and the OpenAIs? I would say absolutely," Chris Canal, chief executive of the AI safety evaluator EquiStamp, told The Verge. "One of the claims is that they're going to permeate every aspect of our lives and be able to do anything that we could do online. If there are companies that are actively working to make that not happen, there are just fewer useful things the agents can do."
Not every door is closing. Meta, whose Muse agent is among those affected, argues that "turning away a personal agent means turning away the customer behind it," and says it would rather work with businesses than see them lose that customer. It has partnered with retailers including Walmart, Gap, Sephora and Best Buy so Muse can shop on their sites, and Shopify has opened its network of storefronts to AI agents. Even where no ban exists, agents trip the same anti-bot safeguards built to catch scammers, and users report account flags and failed checkouts even on sites that welcome them.
Cloudflare, whose tools help sites screen automated visitors, has made blocking easier, and Canal warns that the friction could spill onto everyone, with ordinary shoppers asked to jump through hoops to prove they are human. The agentic web, in other words, is not arriving as an open frontier. It is being negotiated site by site and license by license, and for now several of the biggest doors on the internet remain closed.