The pitch for AI agents is simple: software that can do anything you can do online, from booking a dinner reservation to buying a pair of shoes. In recent weeks, some of the biggest websites on the internet have started answering that pitch with a condition, a price tag, or a closed door, The Verge reported on Saturday.
Users of services like Meta’s Muse, xAI’s Grok, OpenAI’s Dots and Instinct have reported blocks that stop their agents from completing tasks across major sites. Several companies have now confirmed, on the record, that they are keeping agents out — each for its own reason.
Pay up, book elsewhere, or stay out of the account
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. The financial industry is drawing the hardest line: JPMorgan Chase and Charles Schwab have confirmed they are blocking agents from accessing customer accounts, citing security concerns.
Schwab chief executive Rick Wurster explained the position in an interview with Bloomberg, calling out Muse by name: “We don’t want someone just giving their password to Muse, and letting Muse come in and take action on that client’s behalf. We can’t control it, the client can’t control it, something could go wrong and who’s liable? For the protection of the client and for the protection of us, we’re not allowing it.”
The list may be longer than companies admit. Some users say sites that previously worked with their agent are no longer accessible, including Cars.com and Amtrak — a pattern The Verge notes could mean other companies are quietly blocking agents. Neither company is reported to have confirmed a deliberate policy.
Amazon wrote the playbook a year ago
Amazon banned Perplexity’s agentic shopping feature last year and sued the startup, saying it had “repeatedly requested” that Perplexity stop. Companies that buy items on behalf of customers, Amazon argued, “should operate openly and respect service provider decisions whether or not to participate.”
Perplexity answered that Amazon does not want agents roaming its platform because agents do not put human eyeballs in front of its ads. It likened its agent to an extension of the user, arguing Amazon “does not believe in your right to hire labor, to have an assistant or an employee acting on your behalf.” A judge initially ordered Perplexity to stop, but an appeals court overturned that ruling in August. Amazon has since filed an amended complaint.
The newer agents are hitting the same wall. Amazon has told Meta that Muse’s “continued access” violates its site policies, and there are reports of Grok, Instinct’s agent and Dots being unable to use Amazon as well.
The counter-move: deals instead of doors
Meta’s answer, in a post earlier this month, is to negotiate rather than force its way in. “Turning away a personal agent means turning away the customer behind it,” the company said. “We’d rather work with businesses to address the underlying concerns than see them lose that customer.” Meta 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.
That split — licensed access on one side, blocks on the other — is becoming the shape of the agentic web. “Does this impact the perceived value of the Anthropics and the OpenAIs? I would say absolutely,” Chris Canal, chief executive of AI safety evaluator EquiStamp, told The Verge. “If there are companies that are actively working to make that not happen, there are just fewer useful things the agents can do.”
Some of the wall is not aimed at agents at all
Not every failure is a policy decision. Agents also trip the anti-bot systems that blanket the modern web. Reports on Reddit suggest Google has blocked some accounts that use Muse over “suspicious activity,” and other users have reported checkout problems on Walmart despite its Meta partnership. “There are all kinds of things that are set up not even to block agents, but to block in general bots or programs that are trying to scam or hack,” Canal said. “These AI agents trigger those same safeguards.”
Cloudflare, which introduced a one-click block for AI crawlers in 2025, has begun blocking crawlers that scrape sites for both search indexing and AI training at once. Canal warns the side effects will land on everyone: the web could get more “annoying,” he told The Verge, as sites add hoops “to prove that you’re a human using the computer or your phone.”
The Verge said it reached out to OpenAI and Google for comment and did not immediately hear back, while Instinct declined to comment. For now, the agents can still browse much of the web. What they increasingly cannot do is assume the web has agreed to let them act on it.
