AI · October 10, 2026
Meta, Walmart and Stripe launch open standard for AI agents
Meta, Walmart, Stripe and other companies announced on Tuesday the creation of an open technology standard to regulate how personal artificial intelligence agents interact with businesses. The initiative aims to reduce what OpenAI board president Bret Taylor described as "chaos" in the corporate environment.
The effort is led by Taylor through Sierra, an enterprise AI startup he co-founded after leaving his role as co-CEO of Salesforce. In an interview, Taylor explained that companies need mechanisms to distinguish AI agents from real human users. He said businesses must know when they are dealing with a personal agent and not a real person, warning that any bot not acting on behalf of a person should not have access to services. "It is a kind of chaos until this standard exists," he said.
The initiative comes amid the rapid growth of Muse, the AI personal assistant launched by Meta in September. Muse became the most popular AI personal agent on the market, reaching the top of Apple's App Store and remaining in first place ahead of ChatGPT. However, this growth triggered reactions. Amazon blocked Meta's agents, citing concerns about data scraping from its websites. Amazon also blocked the AI agent from Perplexity and filed a lawsuit against the startup in November, alleging that the company took steps to "hide" its agents and continue collecting data from Amazon's site without authorization. Perplexity called the lawsuit a "harassment tactic."
David Singleton, vice president of engineering and consumer products at Meta Superintelligence Labs and former technology director at Stripe, said the standard was designed to provide visibility and control over what agents do on websites. He noted that for agents to function properly, customers need to share credit card and personal information. "We are setting paths along which we hope personal agents and corporate agents can travel in the future," Singleton said, comparing the proposal to email. "If you think about email, it works very well because it is a standard that everyone can use to communicate, and we believe this will be quite similar."
Taylor compared the new authentication method to logging into other sites using Google or Facebook credentials, a technology he worked on while at Facebook. The mechanism should allow companies to identify bots and track what personal agents are doing in their systems.
Singleton said Muse already has millions of users in the United States and is growing "rapidly." He explained that people use Muse for various tasks such as enrolling children in classes, finding and buying gifts, and scheduling dentist appointments they have been postponing for months. "We need a mechanism for this to happen faster and more calmly."
OpenAI and Anthropic are not part of the group for now. Taylor said he hopes both companies will join and said he would be "very disappointed" if the standard is not adopted by competitors. "The goal of all this is to be an open standard," he said.