Websites Are Closing Their Doors to AI Agents: Yelp, Delta, JPMorgan Chase and Charles Schwab Block Agentic Access
Autonomous AI agents are being locked out of major websites. In October 2026, Yelp, Delta Air Lines, JPMorgan Chase and Charles Schwab all confirmed restrictions on AI agents acting on behalf of users, according to reporting by The Verge's Emma Roth. The blocks mark the first broad, coordinated pushback against the agentic web that OpenAI, Meta, Google and others have been promising.
What happened: sites began blocking AI agents in October 2026
The Verge reported on October 17, 2026 that users of several agent products — Meta's Muse, xAI's Grok, OpenAI's Dots, and Instinct — encountered blocks preventing their agents from completing tasks across top websites. The article, titled "Websites are closing their doors to AI agents," was written by Emma Roth and published at 1:00 PM UTC.
The framing of the piece is direct: OpenAI, Meta and Google want AI agents to do everything for you, but popular services on the web are not letting that happen. Agents are marketed as tools that can scour the internet on a user's behalf, handling tasks from booking a dinner reservation and cleaning out an inbox to buying new shoes.
Which companies are blocking AI agents
Several named companies have confirmed or been reported to block agentic access. The reasons split into two categories: monetization and security.
- Yelp told TechCrunch that it only allows AI agents that pay for its data licensing program.
- Delta Air Lines is preventing AI agents from booking tickets on behalf of users.
- JPMorgan Chase confirmed it is blocking AI agents from accessing accounts.
- Charles Schwab was reported by Bloomberg to be blocking AI agents attempting to move money for customers.
Both JPMorgan Chase and Charles Schwab cited security concerns as the reason for blocking agents. The Verge also noted that some users say sites that previously worked with their AI agent are no longer accessible, including Cars.com and Amtrak — suggesting other companies may be blocking agents quietly without public announcements.
The Amazon vs. Perplexity precedent set the tone
Amazon was already fighting AI agents before the October 2026 wave. According to The Verge, Amazon banned Perplexity's agentic shopping feature and sued the AI startup. Amazon claimed it "repeatedly requested" that Perplexity stop allowing its AI agents to shop on its website, adding that third-party companies that purchase items on behalf of customers "should operate openly and respect service provider decisions whether or not to participate."
Perplexity pushed back in a blog post titled "Bullying is not innovation." The company argued that Amazon does not want AI agents roaming its platform because it would mean fewer human eyeballs in front of Amazon's ads. The Verge noted that this concern is real for the ecommerce giant.
Why websites block agents: ads, upsells and security
The core conflict is economic. Many consumer-facing websites earn a large share of revenue from advertising and upsells — elements that AI agents can be instructed to ignore and that agents do not interact with the way a human shopper does.
A widely upvoted Hacker News comment on the story summarized this dynamic:
"A lot of these companies that are closing their doors make a large portion of their revenue based off of ads and upsells which agents can easily be instructed to ignore, and don't interact with like a human. Another more blunt way to say it is that people can instruct their agents to look out for their best interests... and companies hate that."
That commenter predicted the outcome will be a long attrition war: "People will work around it similar to how we've worked around with adblockers for years, and it will just become a game of attrition as these companies work to preserve their moats until they find a way to preserve this revenue stream off of people's agents."
Agents flatten comparison shopping, which threatens retail moats
A second theme in the discussion is that agents undermine the friction that keeps customers loyal to a single retailer. One commenter framed Amazon's business model this way:
"Companies don't want it to be easy to cross shop their competition, the whole premise of Amazon is that it's so easy and simple you don't even check the competition. AI agents are leveling the playing field and helping customers find the best deals, so of course they're being banned."
Another commenter extended the argument to pricing structures more broadly, noting that without a convenience moat, commerce becomes more adversarial through "membership pricing, surveillance pricing, loyalty perks, varied products to avoid price comparison (like mattresses and Costco)." That commenter added a counterpoint: if AI is genuinely superhuman in capability, it may still win for the customer through "aggressive price comparison, alternatives discovery, negotiation, group buying."
Not everyone agrees that agent-driven price competition helps consumers. One commenter argued that past tech disruptions have not passed savings through:
"People who rally against Amazon in this topic assume that the 'savings' from much open price competition immediately come back to the consumer or the manufacturer. In my experience, with previous such tech improvement such as 'uber disrupting taxies', this has never been the case."
Bot detection is catching humans too
The blocks are not only hitting agents. One commenter described being locked out of services as a human user because of third-party bot protection:
"I'm unable to register a GitHub account because the third-party bot protection service they use, called DataDome, thinks I am a bot just based on my IP and probably my extension combination, both in Firefox and Chromium. Datadome is also preventing me from accessing some regional sites of Leroy Merlin, Idealista, and a few other major websites in Portugal and Spain."
The same commenter said DataDome had previously told them that "fighting false-positives is their biggest priority," but that when they contacted the company again recently, the response changed to: "our support channel is for customers only and this ticket will be closed." The commenter added that DataDome's customers often have no idea the service is interfering with their own end users.
The pattern is not new: a little automation is good, a lot is blocked
One of the most detailed comments placed the agent-blocking wave in a longer historical cycle of web openness and enclosure:
"Email was awesome when it first became widespread... Then lots of people (and companies) started cold-emailing, and it stopped working. Social media is broken if there's nobody there, is great with a medium-sized community, and is broken if everyone is there. Ad-supported websites can offer all kinds of valuable free services. Up until either (1) they become so ad-infested that they're no longer useful, or (2) all the users block ads."
The same commenter traced the sequence that led to the current state: websites offered APIs for programmatic access, users scripted alternate interfaces that skipped ads and engagement funnels, and "everything got locked up behind captchas and accounts and API keys with limited access." Their conclusion: "Agents don't change the underlying structure at all. They just magnify and accelerate parts of the system, and will therefore magnify and accelerate the reactive and defensive responses."
Proposed fixes: micropayments, machine-readable site versions and the private web
Hacker News commenters floated several possible resolutions to the standoff.
Micropayments for agent access. One commenter proposed that sites "sell a 'web search' subscription add-on that gets distributed to the sites your agents utilize," describing it as "a form of micropayment system that the internet has desperately needed for over a decade now." Another pointed to Cloudflare's new service offering micropayments to agents, predicting it "is going to win and convert the ad ridden web to the ad-free web."
Separate machine-readable versions of sites. One commenter predicted that "websites will have a separate xml/json/yaml version that agents can peruse without all the ads/js/etc."
A retreat to private spaces. Another commenter forecast the opposite of openness: "In the wake will be the rise of private websites. No social media, no public profiles. Your own private circles."
A directory of agent-friendly shops. One commenter said they built a site, shopsforbots.com, listing sites that do not block bots and let them shop without getting blocked.
The asymmetry complaint: sites block agents while scraping users
A recurring grievance in the discussion is that the same companies blocking agents are themselves aggressive scrapers. One commenter described the double standard:
"It's also a 'you can't scrape us but we can scrape you' mentality these companies are getting into. It really sucks for small businesses. Their websites get absolutely hammered and they can't afford or don't know how to stop it. Cloudflare ends up winning as they're the easiest solution."
Another commenter noted the irony of Meta's position specifically, saying Meta is "one of the most aggressive companies against any sort of automation, perhaps after LinkedIn," and describing being blocked from LinkedIn for letting an agent toggle off 77 email notification settings.
What the standoff means for the agentic web
The Verge's reporting and the Hacker News discussion converge on one conclusion: the agentic web will not arrive uniformly. Access will be negotiated site by site, with financial services and travel leading the blocklist on security grounds and ad-supported commerce leading it on revenue grounds.
One commenter summed up the consumer-side response: "I'll read those arguments and buy those things my agents can find on the open web; which is plenty... Want to sell the greatest widget since the invention of sliced bread? Well, I never found it. Sorry, I bought inferior widgets from your competitor instead."
Another offered a bleaker view of where the open web ends up: "The openweb will go the way of usenet. Bots talking to bots, fetching information for a higher tool."
Note: The Verge article is partially paywalled, so this summary reflects the publicly available portion of the reporting plus the Hacker News discussion thread.