Very proud of what we've build so far this year on @Shopify
Is your @Shopify store prepared for AI agents? I recorded this video about @joinagentready to explain the app I've been working on for 8+ months and would appreciate any feedback or sharing.
I've put a lot into this and love what it's turning into.
1
24
Three things that get collapsed into one number, and should not be:
Cited. The answer used or linked your page.
Named. The answer said your business out loud.
Recommended. The answer put you forward as the choice.
1
3
They arrive in that order and they arrive apart. A page can support an answer that credits someone else entirely, or no publisher at all.
That is still real retrieval evidence and worth having. It is not "AI recommends us," and each one needs its own denominator before it means anything.
joinagentready.com/blog/can-…
4
Before an AI shopping agent can compare your price to a competitor's, it has to decide the two of you are selling the same product.
That step has a name and almost no merchant is watching it.
1
17
Clustering. Offers that represent one underlying product get grouped so merchants can be compared against each other rather than treated as unique listings.
What helps: titles naming the product, clean variant mapping, GTINs, a category that matches how shoppers ask. What hurts: near-duplicate listings and an invented internal taxonomy.
1
12
Where gaps exist, the catalog infers. Those guesses are exactly where you stop controlling your own identity.
Miss the cluster and you never reach the comparison at all. You do not lose the buy box, you are just not in it. cc @ShopifyDevs
joinagentready.com/blog/how-…
16
Stale availability is worse than a wrong label.
A product an assistant cannot confirm is in stock is a product it will not put in front of a shopper. So it does not show your store with a sold-out badge. It leaves you out of the answer and recommends someone else.
Restocks and multi-location inventory are where this bites hardest on @Shopify.
joinagentready.com/blog/why-…
5
Your product page shows 4.8 stars from 400 customers.
An AI assistant reads the same page and finds nothing at all.
1
5
Most review widgets paint their content with JavaScript after the page loads. An assistant fetching the HTML gets the empty container and none of the text.
The stars are real. They are just not in the response anyone machine-readable receives.
1
4
Takes about a minute to check. View source on a product page, search for aggregateRating and for a phrase from your most recent review.
Two hits and both jobs are being done. Stars on screen, nothing in source, and you have been buying display.
joinagentready.com/blog/revi…
5
Two user agents worth telling apart in your logs:
GPTBot is broad crawling.
ChatGPT-User means a real person, in a real conversation, caused that fetch.
The second is much closer to buying intent.
1
9
Heavy GPTBot with almost no ChatGPT-User is a specific diagnosis rather than a mystery. You are being indexed and not surfaced.
The gap there is usually confidence. An assistant will crawl almost anything and recommend only what it can verify.
joinagentready.com/blog/meas…
3
"Ships to Canada" is not a preference an assistant weighs against your price.
It is a filter. Fail it and you are removed from that shopper's results before ranking even begins.
Which makes shipping configuration a discovery problem, not just a fulfillment one. A market you actually serve but have no rates configured for reads as a market you do not serve. Worth an audit of your @Shopify shipping profiles against the markets you sell to.
joinagentready.com/blog/retu…
1
7
A password page on your Shopify store removes your entire catalog from AI shopping.
Not some of it. All of it.
1
9
Catalog inclusion is automatic and invisible. Eligible products flow into the index ChatGPT, Gemini, Copilot and the Shop app read. Ineligible ones quietly do not, and nothing in your admin tells you which camp you are in.
1
15
Per product you need a title, at least one image, a price above zero, publication to the Online Store, and no hidden-from-search-engines flag.
Bulk imports and visibility settings flipped during a sale cause most real failures. Neither one announces itself.
joinagentready.com/blog/shop…
8
A Shopify product is not one machine-readable record when its options differ. Price, availability, identifiers, and images can belong to a specific variant. Preserve that detail instead of flattening the catalog. joinagentready.com/features/…
6
Theme updates can break schema invisibly. AgentReady records publishes, checks, drift, and recovery. joinagentready.com/features/…
A storefront can look fine while its machine-readable facts are incomplete. AgentReady’s free checker reads public pages and reports structured-data and discovery gaps. No install and no ranking promise.
joinagentready.com/tools/sho…
6
Two Product schema blocks can disagree about price, availability, or ratings. AgentReady checks for competing markup so a merchant can keep one authoritative source. Clearer data, no ranking promise.
joinagentready.com/features/…
1
7
Complete the entire AgentReady onboarding within @shopify sidekick, quickly accessing all relevant structured data information anywhere in your Shopify admin.
10


