On October 6, 2026, Sierra and Meta announced the Personal Agent Protocol (PAP), an open standard designed to govern how personal AI agents authenticate and interact with businesses. As the industry shifts toward agentic commerce, the lack of a unified handshake between consumer-facing agents and enterprise systems has created significant friction. By building on OAuth, the protocol aims to standardize how agents identify themselves, declare their intent, and execute tasks on behalf of users.
The technical architecture of PAP focuses on session-based authorization. It allows users to grant agents specific permissions—ranging from read-only access to write capabilities—that persist across different digital channels. Businesses can integrate these agents through their existing websites, APIs like MCP or OpenAPI, or by routing them to their own proprietary agents. As Sierra co-founder Bret Taylor noted to CNBC, “It is kind of chaos until such a standard exists.”
The initial coalition backing the protocol includes Stripe, Shopify, Walmart, Genesys, Instinct, and Rocket. These partners bring a mix of payment infrastructure, retail scale, and customer service expertise. For merchants, the value proposition is clear: establishing a trusted, secure channel for AI interaction. Tony Bates, Chairman & CEO of Genesys, emphasized this in a company blog post: “Personal AI is creating a new front door to the enterprise. Brands need a trusted way to know who an AI agent represents, what it’s authorized to do, its intent, and how to work with it securely.”
However, the protocol’s long-term viability faces a significant hurdle: the absence of the industry’s most influential players. Amazon, OpenAI, and Anthropic are not currently part of the founding group. This split is the story of the week. Amazon has already demonstrated a defensive posture, having blocked Meta’s agents and initiated legal action against Perplexity over website scraping concerns. Without the participation of these major model providers and the world’s largest e-commerce platform, the “universal” nature of the standard remains aspirational.
The challenge is compounded by a deep-seated trust deficit. According to data cited in recent WSJ coverage, only 3% of US adults trust AI agents to complete purchases, while one-third of active AI users state they would never allow an agent to buy on their behalf. Merchants are similarly divided; some are actively welcoming agents, while others are browsing but blocking purchases, or opting to block agent traffic entirely. This environment makes the standardization of “agent posture” a critical, yet difficult, objective.
It is important to distinguish PAP from the growing list of payment-focused protocols. While efforts like Mastercard trust scoring focus on the settlement layer, PAP governs the interaction layer that sits above it. It is designed to handle the “who” and the “what” of the transaction, leaving the “how” of the payment to existing financial rails. David Singleton, VP at Meta Superintelligence Labs and former CTO of Stripe, compared the ambition of the protocol to email, telling CNBC, “We’re defining rails that we hope personal agents and business agents can run over for the future.”
The landscape is currently crowded and consolidating. This follows a period of rapid experimentation, including Shopify’s default-on agent checkout for Meta’s Muse—which has seen rapid adoption since its September 2026 launch—and TikTok’s Shopping Assistant and Buy Direct features.
For the Personal Agent Protocol to succeed, it must move beyond its current partner list. Taylor has stated he expects OpenAI and Anthropic to participate eventually, but their current absence leaves a vacuum. As the industry builds out this trust stack, the success of PAP will depend less on the elegance of its OAuth implementation and more on whether the giants of the AI and retail sectors decide that a common front door is preferable to the current state of fragmented, defensive walls.
Note: Trust data regarding consumer sentiment toward AI agents is sourced from PartnerCentric via WSJ coverage; these figures have not been re-verified against primary reports in this cycle.