Learn / Is my store ready
Do I need to do anything for AI shopping agents?
Most of what makes a store work for AI shopping agents is data hygiene you would want anyway: prices and availability as structured data, terms as text, a checkout that works signed out on a phone. Almost nothing requires adopting a protocol or building anything new. The one thing worth doing now is making sure your bot rules are a decision rather than an inheritance.
Probably less than you fear, and it is not nothing. Most of what makes a store work for AI shopping agents is data hygiene you would want regardless, and the small number of things that are genuinely urgent are urgent because they default against you if ignored.
The one thing that is actually urgent
Make your bot rules a decision. Right now they are probably an inheritance: a robots.txt rule written to keep your content out of training data, and a CDN category setting somebody accepted once.
This is urgent not because agentic shopping is large but because the decision is being made either way, and being made by a default you have not read. Blocking is a legitimate choice. Blocking by accident, discovered a year later, is not a choice at all. See robots blocks agent.
The boring work that pays for itself
Everything in this group improves your store for search, for ads and for ordinary human shoppers as well, which is why it is worth doing before you have decided how big agentic commerce will be.
- Price and availability as product structured data rather than rendered by script.
- Shipping cost and returns terms as text, published before checkout.
- A sitemap that lists products, and a product feed that is current.
- A checkout that reaches the payment step signed out, on a phone, in a cold session.
The full version is the agentic readiness checklist.
What you can safely defer
Adopting a commerce protocol directly. Building an interface specifically for agents. Anything described as an agent optimisation that does not also make a fact checkable. These are all reasonable to leave until the volume justifies them, and none of them substitutes for the fundamentals.
How to decide the size of the bet
Measure before you commit. How much of your traffic comes from AI agents covers what you can measure and what the measurement is worth, and it will usually give you a floor rather than an answer, which is enough to size the work.
If you need to make the case to somebody else, how to explain agent readiness to your leadership team is written to be forwarded.
Questions
Do I have to adopt one of the commerce protocols?
Almost certainly not yet, and probably not directly. The protocols are largely implemented by platforms rather than by individual merchants, and none of them replaces having readable product data. If your fundamentals are missing, a protocol will not compensate.
Can I just wait and see?
You can wait on the optional work. You cannot wait on the bot rules, because that decision gets made for you by defaults at your CDN and by rules already in your robots.txt. Waiting there is choosing, without knowing what you chose.
Is any of this wasted if agentic shopping does not take off?
Very little. Structured product data, published policies, a working guest checkout and an accurate feed all pay for themselves in search, in ads and in ordinary conversion. That is the main argument for doing the boring parts first.