Free Intensive Nov 21–22Free 2-day hands-on Vibecoding Intensive on Nov 21–22 at 16:00 MSK (Sat–Sun) · Live Cohort 3 starts Nov 23Join Free →Cohort 3 (Nov 23)
ReleaseAPI / Product · ActionableTNW→ original

Sierra Releases Poppy 1.0: Shopify Enables Agent Protocol Across All Stores

When a personal AI agent lands on a website, the company needs to know who it is and what it is allowed to do: check an order or book a flight. Sierra has released Poppy 1.0 for exactly this purpose. The customer logs in through a standard login and chooses what the agent can access. Shopify has enabled it across all stores, Meta's Muse agent looks for Poppy on every site, and there are more than 100 partners. Developers should take a closer look, but the specification is still a draft.

AI-processed from TNW; edited by Hamidun News
Sierra Releases Poppy 1.0: Shopify Enables Agent Protocol Across All Stores
Source: TNW. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceTNW ↗
Editor-in-Chief
Key metric #1Shopify has enabled Poppy for all stores, so merchants don't need to configure anything
Key metric #2Mercury lets agents check balances and initiate payments under the rules of the account
Turkish Airlinesthe agent searches for business-class upgrades, and the customer confirms the payment

Personal AI agents are gradually learning to handle routine tasks for people: checking orders, searching for tickets, paying bills. But websites still have no shared way to tell who has come to them. To a server, an ordinary bot, a parsing script, and an agent acting on behalf of a specific customer look almost identical. Sierra, a startup that builds AI for customer support and is led by former Salesforce co-chair Bret Taylor, aims to close this gap with a standard called the Personal Agent Protocol, or Poppy. Version 1.0 is now out, and the standard already has heavyweight backers.

What Poppy is and how it works

To explain it with an everyday example, Poppy works like a visitor badge at a checkpoint. An agent arrives at a site and identifies itself. In response, the company states who it is dealing with and what that agent is permitted to do, for example check an order status or book a flight. The customer signs in through standard logins, as they are used to, and decides what their agent will get access to.

An important detail: Poppy covers only the moment of introduction and permissions. Everything related to the cart and payment is still handled by Shopify's own protocol for purchases. This creates a division of labor: one layer answers "who are you and what are you allowed to do," and the other answers "how to complete the purchase." In addition, for some actions the agent is required to ask the buyer rather than decide on its own.

Who has already signed on

The main news is not the specification itself but who has backed it. Shopify has enabled Poppy for all stores on its platform, and store owners don't need to do anything. Meta's Muse agent now looks for Poppy first on any site. Another 60 companies have joined the design partner program, bringing the total to more than 100. Among them are Delta Air Lines, Intuit, Salesforce, and Zoom.

There are also early production deployments. The banking service Mercury lets agents check balances and initiate payments, but payments still pass through the confirmation rules of the account itself. Turkish Airlines lets agents check fares and search for business-class upgrade options, and the customer confirms the purchase before payment. In both cases the shared principle is visible: the agent gets the right to look and prepare an action, while the final decision about money stays with the human or with existing rules.

Limitations and what to do in practice

The "1.0" label sounds solid, but the specification is still marked as a draft. Sierra explicitly allows changes that could break integrations that are already configured. For teams planning to build production systems on Poppy, this means: plan for a compatibility layer and don't scatter protocol details throughout your code.

Poppy is not the only attempt to bring order to the world of AI buyers. Visa and Mastercard are also developing their own mechanisms for payments made by agents, and Amazon is building its strategy around how shoppers will use AI. It is still unclear whether these approaches will converge on one standard or whether the market will split into several camps.

The practical takeaway differs by role. If you are building an agent, it makes sense to teach it in advance to look for Poppy on a site first, as Muse does. If you run an online store outside Shopify or a service with a personal account, it is time to consider which actions you are willing to entrust to a third-party agent and which should require human confirmation. In any case, keep an eye on how the draft changes: support from Shopify and Meta gives Poppy a good chance of becoming a base layer, but the final rules have not yet been written.

Story

Key Questions Answered

What Poppy is and how it works?

To explain it with an everyday example, Poppy works like a visitor badge at a checkpoint. An agent arrives at a site and identifies itself. In response, the company states who it is dealing with and what that agent is permitted to do, for example check an order status or book a flight. The customer signs in through standard logins, as they are used to, and decides what their agent will get access to. An important detail: Poppy covers only the moment of introduction and permissions.

Who has already signed on?

The main news is not the specification itself but who has backed it. Shopify has enabled Poppy for all stores on its platform, and store owners don't need to do anything. Meta's Muse agent now looks for Poppy first on any site. Another 60 companies have joined the design partner program, bringing the total to more than 100. Among them are Delta Air Lines, Intuit, Salesforce, and Zoom. There are also early production deployments.

What are the key details on Limitations and what to do in practice?

The "1.0" label sounds solid, but the specification is still marked as a draft. Sierra explicitly allows changes that could break integrations that are already configured. For teams planning to build production systems on Poppy, this means: plan for a compatibility layer and don't scatter protocol details throughout your code. Poppy is not the only attempt to bring order to the world of AI buyers.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Private AI Builders Community · Vibe Club

Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews

Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.

Join Vibe Club →Explore Hamidun Academy
What do you think?
Loading comments…