Consumer AI agents, such as Meta's Muse, Instinct and ChatGPT's Dots, represent a new wave of AI that can complete tasks on a user's behalf, from booking flights to ordering groceries. But users are finding that the websites these agents are designed to navigate often block them from finishing the job.
Amazon recently began blocking Meta's Muse agent from its retail site, preventing the agent from browsing its catalog or making purchases. The problem is not limited to Amazon. Complaints across social media indicate agents are frequently running into similar blocks, whether intentional, like Amazon's, or as a result of anti-bot measures designed to stop spam and malicious activity.
What is unclear to users is whether a block is accidental, leaving them frustrated with both the retailer and the agent. TechCrunch has seen a handful of complaints claiming Muse could not complete a purchase on Walmart's website, a problem NBC reported Walmart had faced in its own tests of the AI agent.
A Walmart spokesperson said the blocks were not intentional. The company, which partnered with Muse as announced at Meta's Connect developer conference in September, said it wants to be where its customers are, including through AI agent experiences. The issue appears to be that the website presents a button to verify the visitor is human, and if that check is interrupted, the agent gets booted out.
The episode suggests the technology currently used to verify humans and protect websites from bots may not fit the agent-first web. To address this, industry partners including Meta, Walmart, Stripe, Sierra, Genesys, Rocket, NiCE and Decagon have begun working on an open standard for how AI agents communicate with businesses, helping to separate good bots acting on behalf of users from bad ones.