AI Agents Hit a Wall: Websites Are Blocking Them
Original: The next hurdle for AI agents: getting websites to let them in
Why This Matters
Agent commerce can't scale if websites can't reliably distinguish authorized AI from spam bots.
Consumer AI agents like Meta's Muse and ChatGPT's Dots are being blocked by retail websites when trying to complete purchases. Amazon intentionally blocked Muse; Walmart blocks appear accidental, tied to human-verification systems.
Personal AI agents—Meta's Muse, ChatGPT's Dots, and others—can now book flights, order groceries, and make purchases without the user lifting a finger. The catch: many websites are stopping them cold. Amazon made headlines by deliberately blocking Muse from browsing or buying on its retail platform. But not every block is intentional. Walmart, which is actually a Meta partner announced at Connect in September, has been accidentally blocking Muse users through standard CAPTCHA-style human verification prompts. If that check fails mid-session, the agent gets kicked out. Users can't always tell whether a block is policy or a technical glitch—leaving them frustrated with both the retailer and the agent. NBC previously reported Walmart had encountered this issue in its own internal tests. The root problem: anti-bot infrastructure built to stop spam was never designed to distinguish between malicious crawlers and legitimate agents acting on a user's behalf. To tackle this, Meta, Walmart, Stripe, Sierra, Genesys, Rocket, NiCE, and Decagon have announced work on an open protocol specifically for agent-to-agent commerce communication—a way to signal that a bot is authorized by a real consumer, not a bad actor.