Websites are closing their doors to AI agents

Big tech is pushing autonomous AI agents that act on users’ behalf - booking reservations, shopping, managing email - but many major websites are actively blocking those agents. Retailers, airlines, and financial institutions are restricting agent access: Yelp requires paid data licenses, Delta prevents ticket booking, and banks like JPMorgan Chase and Charles Schwab block account access citing security. Some sites that once worked with agents, including Cars.com and Amtrak, are now unreachable, suggesting widespread, sometimes quiet, blocking. Amazon fought Perplexity over an agentic shopping feature, sought an injunction and argued agents reduce human eyeballs on ads; a judge temporarily barred Perplexity but an appeals court later overturned that ruling, and Amazon filed an amended complaint. Meta’s Muse and other agents run into similar blocks, hurting the agents’ promised usefulness.

Web defenses amplify the problem. Anti-bot systems and services like Cloudflare have added controls - Cloudflare implemented a one-click block for AI crawlers and began blocking scrapers used for both indexing and training - and those safeguards flag agents as suspicious, producing account lockouts and failed checkouts. Meta has responded by partnering with selected retailers and Shopify opened its storefront network to agents, but broader blocking and stricter bot checks will force more sites to add human-verification steps and reduce where agents can reliably operate.