By Steve Merrill, Founder of WRKNG Digital | August 15, 2026
How do you audit a Shopify store for AI visibility without any tools?
You run seven checks in a browser: read your product structured data with view-source, confirm it in Google's Rich Results Test, ask ChatGPT and Perplexity and Google AI Mode about your products and category, check your sitemap lastmod dates, look for an llms.txt file, test your FAQ against real buyer questions, and verify your brand name is consistent everywhere. Score each pass or fail. That's the whole audit.
No software. No login walls. Just you, a browser, and about 30 minutes. I've run this exact sequence on dozens of stores, and it finds the same gaps a $500-a-month platform finds.
Step 1: Can AI read your product structured data?
Open one product page. Right-click and pick View Source, or open Inspect and search the HTML for "application/ld+json." You're looking for a block with "@type": "Product" that includes name, offers, price, availability, and brand.
If it's there, good. If the price and availability are missing or wrong, AI can't trust your listing. Now copy the URL into Google's Rich Results Test. It's free, and it tells you exactly what a machine reads on that page. The Schema.org Product type is the reference for what each field should hold.
Most Shopify themes output basic Product schema by default. Most also leave holes. I audited a store last month where every product had schema, but the availability field said "InStock" on items that had been sold out for weeks. AI reads that field literally. Recommend a dead product once and the assistant stops trusting your feed.
Pass if view-source shows Product schema with price and availability, and the Rich Results Test parses it clean.
Step 2: Does AI name your store when you ask about your own products?
Pick a real question a buyer asks before finding your product. Not your brand name. The actual problem they type. Then ask it in ChatGPT, in Perplexity, and in Google AI Mode, one at a time.
Watch what comes back. Does your store get named? Cited with a link? Or does the answer list five competitors and skip you entirely?
Perplexity is the most useful of the three here because it shows its sources inline, so you can see exactly which pages it pulled. If a competitor's page shows up and yours doesn't, that's your gap staring back at you.
Pass if at least one assistant names or cites your store for a real product question.
Step 3: Does your brand show up for the category, not just the exact product?
Broaden the question. Ask each assistant to recommend the best options in your category. "What are the best organic cotton baby onesies?" "Which brands make durable canvas work aprons?"
This is the harder test. You're competing against every brand the model knows, and it only names a handful. If you never appear, the model doesn't associate your brand with your category yet.
Here's the thing. This is the test most store owners fail, and it's the one that matters most for growth. Product-level visibility captures people who already know roughly what they want. Category-level visibility captures people at the start of the buying decision, when they're asking AI who to even consider.
Pass if your brand appears in at least one category recommendation across the three assistants.
Step 4: Are your sitemap dates telling AI you're active?
Go to yourstore.com/sitemap.xml in your browser. Shopify generates this automatically. Look at the <lastmod> dates next to your product and page URLs.
Those dates tell crawlers when a page last changed. Fresh, accurate lastmod dates signal an active store worth re-crawling. Dates frozen a year in the past signal a store nobody's touched.
Shopify updates lastmod when you actually edit a page, which is correct behavior per Shopify's sitemap documentation. So if your dates are all stale, the real problem is that your content hasn't changed in a year. That's the signal to fix, not the date itself.
Pass if your key pages show lastmod dates from the last few months that reflect real edits.
Step 5: Do you have an llms.txt file?
Type yourstore.com/llms.txt into your browser. Most stores get a 404. Blank page.
llms.txt is a proposed plain-text file that sits at your domain root and points AI crawlers to your most important pages, the way a table of contents would. It's an emerging convention, not an official standard from OpenAI or Google, and Shopify has no native setting for it yet.
So a 404 here isn't a penalty. Nobody gets dinged for missing it. But adding one is cheap, and it's a small signal that you're thinking about how machines read your store. I treat this as a bonus point, not a dealbreaker.
Pass if a file loads at /llms.txt and lists your core pages. Partial credit for knowing it's missing.
Step 6: Does your content actually answer the question buyers ask?
Pull up your FAQ page and any definition or guide content. Read each answer against the exact question a buyer would type. Is the answer direct and quotable in the first two sentences? Or does it bury the payoff under three paragraphs of setup?
AI assistants lift short, direct answers. A 40-to-60-word response that answers the question head-on is what gets pulled into an AI answer. A rambling intro that circles the point gets skipped.
Read it out loud. If you'd be annoyed hearing that answer on a phone call, AI won't use it either. This step has no tool. It's just you being honest about whether your content earns the citation.
Pass if your top questions get a direct, quotable answer in the first two sentences.
Step 7: Is your brand the same entity everywhere?
Search your brand name. Look at your homepage, your Google Business listing, your social profiles, and any marketplace pages. Is the name spelled the same way? Is the description consistent? Does everything agree on what you sell and who you are?
AI models build an entity around your brand by matching signals across the web. When your site says "WRKNG Digital," Instagram says "Working Digital," and a directory says "WRKNG Digital LLC," the model isn't sure those are the same company. Confusion splits your authority.
Consistency is boring and it's the cheapest visibility win there is. Same name. Same one-line description. Everywhere.
Pass if your brand name and description match across your site and the top external listings.
How do you score the audit?
Give one point for each step you pass. Steps 1 through 4, 6, and 7 are full points. Step 5 (llms.txt) is a bonus point since it's optional.
Here's how to read your score out of 7:
- 6 to 7: Strong. AI can find, parse, and recommend you. Keep re-running the question tests monthly.
- 4 to 5: Fixable. You have the foundation. The gaps are usually category visibility and content depth.
- 2 to 3: At risk. AI can barely see you. Start with structured data and content that answers real questions.
- 0 to 1: Invisible. AI shoppers are being sent to your competitors right now. This is the emergency bucket.
The score isn't the point. The failed steps are. Each fail is a specific, fixable thing between your store and an AI recommendation.
FAQ
Can I audit my Shopify store for AI visibility without paid tools?
Yes. A browser, view-source, Google's free Rich Results Test, and free access to ChatGPT, Perplexity, and Google AI Mode cover every check that matters. Paid tools speed up tracking at scale, but they don't find anything a careful manual audit misses on a single store.
Does my Shopify store need structured data for AI to recommend it?
Yes. AI assistants read Schema.org Product and Offer markup to pull your price, availability, and brand as facts. Most Shopify themes output basic Product schema, but confirm it with view-source and the Rich Results Test rather than assume it's there.
What is llms.txt and does Shopify support it?
It's a proposed plain-text file at your domain root that points AI crawlers to your most important pages. It's an emerging convention, not an official standard, and Shopify has no native setting for it yet. Missing it won't get you penalized, but adding one is a low-cost signal.
How often should I run an AI visibility audit?
Full audit once a quarter, question tests monthly. AI answers shift as models and their sources update, so a store that gets recommended in March can drop out by June without a single change on your end.
Want us to run this audit on your store and fix every failed step so AI actually recommends you? See how at WRKNG Digital's agentic commerce page.
By Steve Merrill, Founder of WRKNG Digital. August 15, 2026.
