Consumer SI agents like Meta’s Muse and ChatGPT’s Dots are facing a new barrier: websites are blocking them from completing tasks. While these tools aim to automate actions such as booking flights or ordering groceries, users report frequent interruptions when retail and service platforms reject the agent's traffic.
What Happened
Amazon recently blocked Meta’s Muse SI agent from browsing or purchasing from its product catalog. However, the issue extends beyond intentional bans. Users have reported that agents from various providers are often stopped by traditional anti-bot measures, such as human-verification buttons, which can fail if the automated process is interrupted. For instance, complaints on social media indicated that Muse could not complete purchases on Walmart’s site, despite Walmart announcing a partnership with the agent at Meta’s Connect conference in September. A Walmart spokesperson stated the blocks were unintentional.
Other major companies have taken more restrictive stances. Delta Air Lines confirmed it does not currently have a partnership enabling third-party SI agents to book flights, stating that any integration would need to prioritize "security and the customer experience." United Airlines pointed to its Terms of Use, which prohibit the use of robots or automatic devices without prior written permission. Yelp also restricts non-human traffic unless the agent accesses data through its paid licensing program.
Why It Matters
The friction between SI agents and existing web infrastructure highlights a gap in the current digital ecosystem. Traditional security tools, designed to stop spam and malicious scrapers, are inadvertently blocking legitimate agents acting on behalf of human users. This creates confusion for consumers, who may perceive the failure as a flaw in the SI agent rather than a policy decision by the retailer. Cloudflare, a major provider of web security, noted that recent changes to crawler defaults allowed site owners to block SI training while permitting other bots, but it did not have specific data on the impact to personal SI agents. The situation has prompted industry partners, including Meta, Walmart, Stripe, and others, to begin working on an open standard for agent-to-agent communication in online commerce.
The Bottom Line
As SI agents become more prevalent, the web’s existing gatekeeping mechanisms are proving ill-equipped to distinguish between helpful automation and malicious activity. While some companies like Meta are pursuing direct partnerships and new standards to resolve this, many major retailers and service providers remain closed to unauthorized agents, leaving users to navigate a fragmented landscape where access is not guaranteed.