
theverge.com
October 18, 2026
5 min read
53/100
Summary
AI agents from companies including OpenAI, Meta, Google, xAI and Instinct are encountering blocks on major websites as they attempt to shop, book travel, access accounts and complete other tasks for users. Yelp permits agents only through its paid data-licensing program, while Delta Air Lines is preventing them from booking tickets. JPMorgan Chase and Charles Schwab say they block agents from account access over security concerns. Users have also reported agent access failures on Cars.com, Amtrak and Amazon. Amazon previously banned Perplexity’s agentic shopping feature and sued the company, saying Perplexity ignored requests to stop agents from shopping on its site. A judge initially ordered Perplexity to halt the activity, but an appeals court overturned that order in August; Amazon has filed an amended complaint. Meta says denying a personal agent also denies the customer behind it, and it has partnered with Walmart, Gap, Sephora, Best Buy and other retailers to support Muse shopping. Shopify has opened its storefront network to AI agents. Anti-bot systems can also block agents unintentionally: users have reported Google account blocks tied to Muse and Walmart checkout problems. Cloudflare introduced one-click AI-crawler blocking in 2025 and blocks some crawlers used for both search indexing and AI training.
Key Takeaways
What the discussion said
Commenters treated the closures less as a technical fight over bots than as a collision between agent-mediated shopping and an ad-funded web. Many expect people to prefer agents that compare prices, inspect specifications, handle tedious settings, and filter manipulative interfaces. From that view, blocking agents is a defensive move by platforms that profit when customers cannot easily compare alternatives or skip advertising, subscriptions, and upsells. Several readers argued that businesses which hide their catalogues from agents will simply disappear from increasingly automated buying decisions. There was also a practical vision of a parallel machine-readable web: structured feeds or paid APIs through which agents can retrieve reliable information without rendering the human-facing maze of scripts, banners, and tracking. Some saw this as a route toward agent-search optimization and micropayments for publishers. But enthusiasm was checked by the economics. Agents consume services while withholding the attention and ad impressions that finance them, so unrestricted access could destroy the very sites users expect agents to search. Others warned that comparison engines tend to overweight price, while shipping, returns, service, and product quality matter. The thread also noted an ugly immediate consequence: anti-bot systems already lock out legitimate humans, while smaller sites absorb abusive automated traffic or become dependent on large protection vendors.
Where opinion split
The sharp dispute is whether agents liberate consumers or hollow out the web that serves them. Supporters say agents expose better offers, eliminate hostile interfaces, and force sellers to compete on value; critics say they evade the advertising-funded bargain and may reduce choices to a price race while starving useful services of revenue.
Community Sentiment
Positives
Concerns

Wikipedia's AI agent row likely just the beginning of the bot-ocalypse
Apr 6, 2026

What's missing in the 'agentic' story: a well-defined user agent role
Apr 25, 2026

AI Resistance: some recent anti-AI stuff that’s worth discussing
Apr 20, 2026

News publishers limit Internet Archive access due to AI scraping concerns
Feb 14, 2026

AI assistant hacks gym website in first known Australian autonomous cyber attack
Aug 9, 2026