AI agents promise to shop, book travel and handle forms for you. But in autumn 2026, a growing list of big companies started turning them away. If you tried an agent such as Meta’s Muse for a purchase or a bank task and it failed, this is likely why. This guide explains who is blocking agents, the reasons they give and what you can do.
Who Is Blocking AI Agents
The Verge reported on October 17, 2026 that Amazon, Yelp, Delta Air Lines, JPMorgan Chase and Charles Schwab have moved to shut third-party AI agents out of the parts of their sites where real tasks happen. Here is what has been reported:
- Amazon has blocked Meta’s Muse since September 20, about twelve days after Muse launched. Amazon says Meta did not warn it, that the agent does not identify itself as AI while browsing and that it appears to store customer credentials. Meta says Muse cannot see passwords or payment methods.
- Yelp admits only agents whose makers pay for its data licensing program.
- Delta is preventing third-party agents from booking flights.
- JPMorgan Chase and Schwab block agents from accessing accounts for security reasons.
- Bank of America’s chief executive has signaled the bank will require human approval for payments until agents prove more reliable.
Schwab’s chief executive said on Bloomberg TV on October 16 that the broker stops dozens of attempts by major AI agents to reach client money every week.
Why Companies Say They Are Doing It
The reasons fall into a few groups. Banks point to security and the risk of an agent moving money by mistake or being tricked. Retailers worry about who controls the customer relationship, ads and pricing. Data-heavy sites such as Yelp want to be paid when their content is used. Some companies also say agents do not clearly identify themselves, which makes it hard to tell a helpful assistant from a harmful bot.
Is Every Block Intentional?
No. Walmart is one of Meta’s announced partners, yet a user still reported a failed purchase. A Walmart spokesperson said the “I’m not a robot” check can throw an agent off when it is interrupted halfway, so that block was not intended. A separate audit of 51 large US retailers, as summarized in industry newsletters, found 37 blocked shopping agents largely by accident through old bot rules, and none offered a way for a verified agent to complete a purchase. A ban and a malfunction can look the same on screen.
Meta’s Response
Meta argues that turning away a personal agent means turning away the customer behind it. It points to live integrations with Walmart, Best Buy, Gap, Sephora and Wayfair, plus a Shopify default that enables Meta checkout for eligible stores. A group led by Meta, with Walmart, Stripe and several customer-service technology companies, is also writing an open standard for how agents talk to businesses, aimed at e-commerce and meant to separate trusted agents from bad bots.
What This Means for You
- Do not assume an agent can complete every purchase or banking task. Check that the site supports it.
- For money movements, expect to approve the action yourself.
- If an agent fails on a site, try the task manually before blaming the tool.
- Never share bank passwords with any agent unless the bank itself offers a supported connection.
- Watch for standards that let verified agents work safely, since that is where this is heading.
For Website Owners
If you run a site, review your bot rules. Old settings can block helpful agents by accident, while real abuse still needs protection. Decide on purpose which agents you want to allow, and make your policy clear.
Final Thoughts
AI agents are only as useful as the places that let them in. Until common rules for identity, payments and permissions exist, expect a patchwork of blocks, partnerships and paid access.