By Steve Merrill, Founder of WRKNG Digital — June 12, 2026
We've audited over 200 Shopify stores for AI readiness. The average score: 52 out of 100. Stores that score 90 or above are getting 3-4x more recommendation impressions from ChatGPT Shopping, Google AI Overviews, and Perplexity. BFCM is roughly 12 weeks out. The window to close that gap is open right now, and it won't be open in October.
This post explains what AI readiness scoring actually measures, what separates a 52-scoring store from a 92-scoring store, and which changes will move the needle most in the next 30-60 days.
What Does AI Readiness Scoring Actually Measure?
AI readiness scoring evaluates whether your store's product data can be understood and recommended by AI shopping assistants. It's a composite score across four areas, and each one has a different ceiling for how much it can hurt you.
Product feed completeness. Your product feed is the primary data source AI shopping tools use to understand your inventory. Fields like GTIN, brand, material, color variant, and product category are the ones AI systems query against when a buyer asks "find me a black merino wool crew neck under $150." If those fields are empty, your product doesn't match. According to Shopify's own product feed documentation, stores with 99% product attribute completion see 3-4x more visibility in AI-powered recommendations. That's a category-level difference, not a marginal one.
Structured data (schema markup). Schema tells AI models what your pages mean, not just what they say. Product schema, Offer schema, BreadcrumbList, and FAQPage markup all give AI systems confidence when recommending your products. Stores missing this markup are asking AI to guess what category a product belongs to, what price it sells for, whether it's in stock. AI doesn't guess when cleaner alternatives exist.
Technical discoverability. The llms.txt file is a simple text file in your root directory that tells AI crawlers what content is available and where to find it. It's present on fewer than 5% of the Shopify stores I've audited personally. The ones that have it are getting indexed faster and more completely. It takes under 30 minutes to add. Most stores don't have one.
Content specificity. AI shopping assistants answer questions. Your product descriptions need to do the same. "High quality materials and expert craftsmanship" tells an AI system almost nothing. "12oz 100% cotton canvas, brass hardware, fits standard 13-inch laptops" is the kind of data that actually gets matched to buyer queries.
What Does a 45-65 Scoring Store Actually Look Like?
Most stores in this range aren't neglected. They're well-run. The problems are invisible from the front end.
A typical 52-scoring store looks like this: Product photos are great. The site loads fast. Checkout experience is smooth. But the product feed has 30-40% of attribute fields empty. Schema markup was added through a generic Shopify plugin and covers Product and Offer, but nothing else. There's no llms.txt file. Product titles include the name and color but not the material, dimensions, or compatibility. Descriptions were written for a human browsing visually, not for an AI model parsing text in response to a specific question.
Invisible to AI shopping surfaces. A real brand, a real operator, completely passed over because the underlying data isn't there.
I've seen this exact pattern in stores doing $2M to $5M per year. They're doing everything right from a merchandising and operations standpoint. The AI readiness gaps aren't obvious until you pull the feed and run it against a structured audit.
Google Merchant Center's product data specification documents over 50 optional attributes that AI-powered shopping tools now actively use for matching. Most stores submit the required fields and stop there. That's enough to appear in traditional Google Shopping. It's not enough to get surfaced by an AI assistant fielding a specific question from a buyer who knows what they want.
What Does a 90+ Scoring Store Look Like?
The gap is specific, not overwhelming.
A store scoring 92 has product feeds at or near 99% attribute completion. GTIN is present on every product that has one. Material and fabric type are filled in. Product category follows Google's taxonomy rather than custom store categories. Titles are structured the way a human would say the product out loud: "Women's Merino Wool Crew Neck Sweater, Navy Blue, Small." Schema markup covers Product, Offer, BreadcrumbList, FAQPage on relevant pages, and Organization at the site level. The llms.txt file exists, is current, and links to the sitemap.
Product descriptions answer specific questions. The copy tells you exactly what something is made of, how it fits, what it's compatible with, and what makes it different from other options in the same category. That's the content an AI assistant can quote directly when answering a buyer's question. It becomes a source, not just a listing.
These stores also tend to have brand pages, collection pages, and category pages with properly marked up content. AI systems use those pages to build a model of what a store sells and who it sells to. Stores without them look like a pile of products with no surrounding context.
What Are the Highest-Impact Changes to Make in the Next 30-60 Days?
Start with your product feed. Nothing else has this much upside per hour of work.
Pull your feed and audit it against Google's Shopping content best practices and the attribute fields ChatGPT Shopping uses for matching. Material, GTIN, brand, product category using Google's taxonomy, and color are your first pass. For most stores, fixing those five field types alone will move a score from the low 50s to the high 60s or 70s. That's a meaningful jump before you touch anything else.
After the feed, add an llms.txt file. Put it at yourdomain.com/llms.txt. The basic format tells AI crawlers what you allow and where to find your sitemap and important content. There's a growing standard forming around this file, and stores that adopt it early are seeing faster and more complete indexing by AI systems.
Schema markup comes next. If you're running a Shopify schema plugin, audit what it's actually outputting. Open your product page source, search for "application/ld+json," and read what's there. Most generic plugins cover Product and Offer but skip FAQPage, BreadcrumbList, and Organization. Add those manually or upgrade your plugin to cover them. A 30-minute schema audit will usually surface 3-4 quick additions.
Then rewrite your top 20 product descriptions. Not all of them. Just the 20 products you most want appearing in AI recommendations during BFCM. Make them specific. Answer the questions a buyer would ask an AI assistant: What's it made of? How does it fit? What size exactly? What is it compatible with? These descriptions become the source content AI systems pull from when recommending your products to someone who hasn't found you yet.
Why Does the Timing Window Matter for BFCM?
AI channels don't work like paid ads. You can't activate them two weeks before BFCM and expect results.
Feed changes can be picked up fast. ChatGPT Shopping refreshes product feed data approximately every 15 minutes once connected. But the authority that feed builds with AI recommendation systems is cumulative. AI models learn which stores consistently provide complete, accurate, high-quality data. Stores that have been doing this for 60-90 days going into BFCM will have a compounding edge over stores that fixed their feed in October.
Schema and content changes take longer. New structured data typically takes 2-4 weeks to fully index and influence AI recommendation outputs. If you wait until mid-October to start, you may not see the full effect until after BFCM is already over.
I watched this exact dynamic play out with Facebook ads. The brands that were already running ads and had clean data when the algorithm shifted were the ones that compounded away from everyone else. The ones who waited until the pain was too great were two years behind and never fully caught up. The same pattern is forming here. Slow burn right now. Hard wall later.
12 weeks is enough time to close most of the gap between a 52-scoring store and a 90+ scoring store. It's not enough time to start thinking about starting.
Frequently Asked Questions About AI Readiness and BFCM
What does an AI readiness score measure for Shopify stores?
An AI readiness score measures how well your store's product data can be understood and recommended by AI shopping assistants like ChatGPT Shopping, Google AI Overviews, and Perplexity. It covers product feed completeness, schema markup, technical discoverability including llms.txt, and how specifically your product content answers buyer questions.
Why do stores scoring 90+ get more AI recommendations than stores scoring 45-65?
AI shopping assistants use product feed data to match products to buyer queries. Stores with complete attribute data, proper schema markup, and clear product descriptions give AI models enough signal to confidently surface their products. Stores with missing or thin data get skipped in favor of results where the AI can answer the buyer's question with confidence.
What is the single most important AI readiness fix for a Shopify store?
Product feed attribute completeness. Shopify's data shows stores with 99% attribute completion see 3-4x more visibility in AI-powered recommendations. Material, color, brand, GTIN, and product category are where to start. These are the fields AI systems use most when matching products to specific buyer queries.
Does improving AI readiness affect Google search rankings?
Improving structured data and product feed quality does tend to improve Google Shopping performance as well, since both systems use product schema and feed data. The primary goal here is AI-powered recommendation surfaces like ChatGPT Shopping, Perplexity shopping tabs, and Google AI Overviews. Better Google Shopping results are a secondary benefit, not the main target.
How long does it take to see results from AI readiness improvements?
Feed-level changes can be picked up by AI systems like ChatGPT Shopping within hours. OpenAI refreshes product feed data approximately every 15 minutes. Schema and technical changes typically take 2-4 weeks to fully index. Building consistent authority for AI recommendations across multiple platforms takes 60-90 days. That's why acting before BFCM means acting now.
Want to see exactly where your store sits on the AI readiness scale, and get a prioritized action plan before BFCM? Get your audit at wrkngdigital.com/agentic-commerce-landing-page.
