How does the Agent-ready score add up?

Most of it comes from what the agent found when it tried to buy. The rest comes from your product data and your store's setup.

The Agent-ready score is out of 100. Most of it comes from what the agent found when it tried to buy, and the rest from your store's data. Every line says which of the two it comes from.

What it's made of

Agent shopping test, no payment 60

From the agent's purchase test

  • Finds your products 20
  • Gets answers from your store 20
  • Prepares a checkout and picks delivery 20

Product data 25

From your data, as a share of your active products

  • Products Shopify Catalog can show 25

Store setup 15

From your data

  • Store facts confirmed 5
  • Agent data on product pages 4
  • Refund policy agents can read 3
  • Shipping policy agents can read 3

How a line scores

  • The agent's lines. A line that passed gets all its points, one that partly passed half of them, and one that failed none. A line the test couldn't check or didn't reach gets none until a later test checks it. So does a line the test can't cover in your country yet, such as the checkout when the test has no address or phone number for your country, on every test until it can.
  • Product data. The share of your active products that Shopify Catalog can show. It leaves out a product with no category or no image, one with a variant priced at zero, and one that isn't published and searchable. If your store has more than 50 products, this comes from the last check of your whole catalog, which runs when you open the Products or Catalog-wide tab of the Fixes page. Without one, it's your 50 most recently updated products.
  • Store setup. Each line is all or nothing. A policy counts once it has at least 10 words. Agent data on product pages, in its default, also needs an agent test that read one of your product pages and saw what it adds attach to your theme's product data: until one has, it gets none.

What isn't in it

  • No score before the first test. Its biggest part would be a guess.
  • Not payment, because the agent never pays, so it's never tested.
  • Not AI answers. They're reported on their own, as counts. What AI answers measure
  • Not the Shipping and returns block. It helps assistants that read your pages, but it adds no points.

How it moves

On the Proof page, the Tests tab lists your tests, newest first, each with the score it saved. The score on Emtor's home page and on the Fixes page is worked out from your store's data now, so the two can differ. When it's the score your newest test saved, the line under it says how far it moved since the test before. On the Fixes page, the To fix tab lists every fix, the first to make at the top, then problems, then the lines that would add the most points. Emtor's home page shows one of them: the step where the agent stopped, if there is one to do there, else the first in that order.