What changed
Personal Agent Protocol (Poppy) 1.0 released; Shopify enables it on every store by default
When
November 13, 2026 (introduced October 6, draft shared October 9)
Who it affects
All Shopify merchants; Meta Muse users; 100+ design partners incl. Walmart, Stripe, Wayfair
What to do
Open yourstore.com/.well-known/poppy.json, then fix catalog and policy data agents will read

Key takeaways

  • Sierra released Poppy 1.0 on November 13, 2026, and Shopify turned it on for every store with no merchant setup.
  • The file at /.well-known/poppy.json tells agents who you are and what shoppers can let them do; UCP still runs the cart and checkout.
  • Agents repeat your catalog and policies, so clean product data and plain return rules now win orders and prevent returns.

Poppy protocol, from idea to every Shopify store

  1. Sierra and Meta introduce the Personal Agent Protocol

  2. Draft spec shared with design partners

  3. Version 1.0 released; Shopify enables it for every merchant

  4. Meta Muse checks every site for a Poppy file first

  5. Black Friday Week: first holiday with agent discovery on

What Sierra and Shopify announced

On November 13, 2026, Sierra co-founders Bret Taylor and Clay Bavor published version 1.0 of the Personal Agent Protocol, nicknamed Poppy. It is an open standard, developed by Meta and Sierra with design partners, that defines how a shopper's personal AI agent connects with a company. The first public version was introduced on October 6 and a draft was shared on October 9, so it went from idea to 1.0 in about five weeks.

The part that touches every Shopify brand: Sierra says Shopify has enabled Poppy for every merchant on its platform, every Shopify store now automatically publishes a Poppy file, and merchants do not need to do any extra work. Shopify also co-authored a commerce extension that links Poppy to the Universal Commerce Protocol (UCP), which Shopify stores already support by default.

  • Agent side: Meta's Muse now looks for a Poppy file first on every site it visits.
  • Partners: initial partners named are Meta, Genesys, Instinct, NiCE, Rocket, Shopify, Stripe and Walmart; 60 more joined, taking the total past 100, including Affirm, Delta Air Lines, Intuit, Salesforce, Wayfair, Wix and Zoom.
  • Live so far: Sierra names Mercury, Synchrony and Turkish Airlines as going live.

How Poppy works, in plain words

The two protocols split the job. In Sierra's words, UCP handles the shopping itself, finding products, filling a cart and checking out, while Poppy handles what happens before that: who the agent is, what it may do, and whether the shopper has signed in. According to the published spec, the file lives at /.well-known/poppy.json on your domain and lists your organization name and domain, the ways an agent can talk to you, any APIs, and the permissions (OAuth scopes) a shopper can grant.

Each permission carries a hint that says whether an action changes nothing, is reversible or is irreversible, and whether it touches personal data. A shopper signs in through your store and approves some or all of the requested permissions, and a company may grant fewer than requested, never more. That is the useful part for brands: an agent can check order status or a return without scraping your site, and irreversible actions are flagged.

PoppyUCP
Who the agent is
What it may do (scopes)
Shopper sign-in
Find products
Fill a cart
Check out

Per Sierra: UCP handles the shopping, Poppy handles what happens before it.

What the coverage misses

Three things most write-ups skip. First, Shopify flipped this on for you, and neither Sierra's post nor the coverage we read describes a merchant setting to turn it off. Second, the spec's own opt-outs are soft: a company can ask agents not to browse its site, but the spec says the protocol cannot enforce that. Third, the status is disputed. Sierra calls it version 1.0, while the implementation kit and Shopifreaks still call the spec a draft that may change in ways that break existing setups. Plan for change, not a finished standard.

The bigger point for AI agents in ecommerce: an agent can only recommend what it can read. Once discovery is automatic, the store with clean product data, clear variants and plain return rules is the one an agent can sell with confidence.

So what: the money and the risk

The money case is a new, low cost referral path. Worked example, illustrative numbers only: a Shopify brand does 4,000 orders a month and pays about $35 to acquire a new customer through paid social. If agent visits bring even 1% of orders, that is 40 orders a month. If half are new customers who would otherwise have come through paid social, the brand avoids about $700 a month in acquisition cost (20 orders times $35), before counting repeat purchases.

The risk runs the other way. If an agent misreads a size chart or a bundle, the order still ships and then comes back. At an illustrative $12 all-in cost per return, 10 wrong orders a month cost $120 plus the customer. The same data fix that wins agent orders also prevents those returns.

  • Upside: discovery and order help without building anything.
  • Downside: agents repeat whatever your catalog and policies say, errors included.

What Shopify brands should do this week

Black Friday Week traffic is days away. These checks take an afternoon:

  • Ecommerce lead: open yourstore.com/.well-known/poppy.json in a browser and confirm it loads with your real store name and domain. Metric: file returns a page, domain matches.
  • Catalog manager: fix product titles, variant names, GTINs, sizes and stock in Shopify admin, because agents shop through your catalog data. Our ecommerce SEO best practices apply here too. Metric: products with missing GTINs or unnamed variants down to zero.
  • Customer service lead: rewrite the shipping cutoff, returns and holiday exchange pages in short, plain sentences an agent can repeat. Metric: tickets that say the bot or agent told me something wrong.
  • Analytics owner: watch referrers and unattributed sessions weekly so agent traffic does not hide inside direct traffic. Metric: weekly sessions and orders from new or unknown sources.
  • Developer: if you run custom sign-in or headless checkout, test with the Poppy kit's validator in a staging store, since it starts real sessions. Metric: validator checks pass.
  • Amazon sellers with a Shopify store: your Shopify site now welcomes agents by default, so decide whether it becomes your agent friendly channel alongside Amazon. See how Shop MCP agentic checkout fits. Metric: share of revenue from your own site.

Agent ready before Black Friday

  • Load /.well-known/poppy.json and check the domain
  • Fill GTINs, sizes and variant names
  • Rewrite returns and shipping cutoffs plainly
  • Track unknown referrers weekly
  • Test custom sign-in with the validator
  • Decide which channel welcomes agents

When this does not apply

If you do not sell on Shopify, nothing was switched on for you, though you can publish a Poppy file yourself using the kit. Brands selling age restricted or regulated products should wait for clearer rules on what agents may do. And low traffic stores will not see agent orders soon: Shopifreaks notes only a few partners are live, and Shopify's existing agent protocol, not Poppy, still runs carts and checkout.

Where PPC Spider helps

PPC Spider is a marketplace agency for brands selling on Amazon and other marketplaces that also grows DTC stores. Our Shopify DTC agency team cleans catalog data and policy pages so people and agents read them the same way, and our AI search optimization work tracks how assistants describe your brand. Selling on Amazon too? Start with a free Amazon audit.

Sources

Get the next change before your competitors do

We send the marketplace changes that matter each week, with what to do next.

By subscribing you agree to our privacy policy. Unsubscribe any time.
Want this done for your account?

We run these playbooks on client accounts every day. Start with a free audit.

Get a free audit