By Steve Merrill, Founder of WRKNG Digital — August 19, 2026
What is an AI commerce audit for a Shopify store?
An AI commerce audit checks whether AI shopping assistants can read, trust, and recommend your products. You review six things: structured data, product feed fields, llms.txt, content extractability, FAQ format, and entity consistency. Then you score each one and fix the worst first.
I've had a hard time getting store owners to run this. They think their site looks great, so it must be fine. Then we open the hood and the data layer is a mess.
Here's how to run it yourself.
How do you check your structured data?
Run every product page through a schema validator and confirm your Product schema carries price, availability, brand, and aggregateRating. This is the first thing an AI model reads. Broken or missing markup gets your product skipped before anything else matters.
Most Shopify themes ship partial markup. You'll often find Product schema with no rating field and no price field. The model reads a shell and moves on.
Test it with the Schema.org validator and Google's Rich Results Test. Score this area out of 10. Full price, availability, brand, and rating on every product page is a 10. Missing the rating field drops you to a 6. No Product schema at all is a 0.
Fix this before you touch anything else. It's the single biggest failure point I see in audits.
What product feed fields do you need to audit?
Check the same core fields Google Merchant Center has required for years: GTIN, brand, title, price, availability, condition, and product category. AI assistants borrow from this spec. Miss a required field and the product drops out before ranking starts.
The gaps repeat across almost every store:
- Empty GTIN. The barcode field is blank, so the model can't match your product to anything.
- Stale availability. The feed says in stock, the page says sold out.
- Price mismatch between the feed and the live page.
- Vague titles like "Blue Shirt" when it should read "Men's Organic Cotton Oxford, Slim Fit, Navy."
The exact requirements live in the Google Merchant Center product data specification. Score this out of 10 based on how many products pass every required field. Above 90 percent clean is a 9 or 10. Half your catalog with holes is a 5.
Does your Shopify store have an llms.txt file?
An llms.txt file is a plain-text file at your root domain that points AI crawlers to your most important pages. Check yourshop.com/llms.txt in a browser. If you get a 404, you don't have one, and most stores don't.
This is early. The standard is new and adoption is thin. But it costs almost nothing to add, and it's a clean signal to models about what matters on your site. You can read the format at llmstxt.org.
Score this one simple. You have a working llms.txt with your key pages listed, that's a 10. You don't, that's a 0. Easy win, easy points.
Can AI actually read your content?
Content extractability means a crawler can pull your text as real HTML, not trapped in an image or loaded only by JavaScript that bots skip. If the model can't read it, it can't cite it. This trips up a lot of themes that render product details client-side.
Test it fast. Open a product page, right-click, and view the page source. Search for your product description text. If it's there in the raw HTML, you're fine. If it only shows up after the page loads in a browser, you have a problem.
I ran this on a client's store last month. Their whole spec table was baked into a single image. The model saw a picture and nothing else. Blank. We rebuilt it as text and their pages started getting pulled into answers.
Score out of 10 based on how much of your core content sits in readable HTML.
How do you score your FAQ and answer format?
Check whether your pages answer real buyer questions in the first sentence. AI assistants pull short, direct answers. A page that buries the answer three paragraphs down loses to a page that states it up front. This is the second place most stores fail.
Look for two things. Question-style headings that match what people actually ask, and a 40 to 60 word answer right underneath each one. Then add FAQPage schema so the model can grab those answers cleanly. Google's own FAQPage structured data guide shows the format.
Most product pages have zero of this. They describe the product. They never answer the question. Score out of 10. Real questions with direct answers and FAQ schema is a 9. A wall of marketing copy is a 2.
Is your entity data consistent everywhere?
Entity consistency means your brand name, product names, prices, and details match across your site, your feed, your schema, and third-party listings. Conflicting data lowers trust, and a model that can't reconcile your details will pick a store it can.
Spot-check five products. Does the price on the page match the feed, the schema, and any marketplace listing? Is the brand name spelled the same way everywhere? Small mismatches add up.
Score out of 10 based on how many of your checks line up clean.
How do you turn the scores into a fix list?
Add your six scores and read the weak spots. The order to fix is set by impact: structured data first, then feed fields, then FAQ format, then extractability, then entity consistency, then llms.txt. Ranking never gets a chance if the data underneath is broken.
Here's the bottom line: structured data and FAQ content are where most stores bleed points. Fix those two and you jump ahead of nearly every competitor in your category.
A store scoring under 40 out of 60 is invisible to AI shopping right now. That's not a death sentence. That's the opening. There's so few stores doing this correctly, which is exactly why it's a good time to move.
Frequently asked questions
What is an AI commerce audit for Shopify?
It's a check of whether AI shopping assistants can read, trust, and recommend your products. You review structured data, feed fields, llms.txt, content extractability, FAQ format, and entity consistency, then score each and fix the biggest gaps first.
How long does an AI commerce audit take?
A first pass takes two to three hours for a small catalog. Fixing structured data and feed gaps takes longer, usually a few days of work with your theme and feed settings.
What do most Shopify stores fail on?
Structured data and FAQ content. Most stores ship incomplete Product schema with no rating or price field, and their pages bury the answer rather than state it up front.
Do I need an llms.txt file for Shopify?
It won't hurt and it's easy. An llms.txt file points AI crawlers to your best pages in plain text. Adoption is early, but it costs almost nothing to add one.
Want us to run the audit for you?
We audit the structured data, feed, and answer content that get Shopify stores recommended inside AI shopping. If you want your score and a fix list, see how we do it at WRKNG Digital's Agentic Commerce page.
