Most stores I have seen in the past 90 days score between 35 and 65 on AI readiness. The tool will tell you where you sit. It will not tell you what to do with that information once your developer asks the obvious next question.
Adam Finan, a Shopify CSM, ran the new Commerce Readiness Tool against Represent, a British apparel brand, in early April 2026. The brand scored 81 out of 100. They had a solid foundation, but the scan flagged product schema and FAQ schema as priorities. Within 24 hours, their dev team had pushed the fixes. That is what the tool does well. It collapses a four-hour audit into 30 seconds and gives a developer a ranked list to work from. The honest assessment came from Adam Finan, who noted that the results are a guide and a signal, not a checklist to blindly execute. That distinction matters more than the score itself.
Here is the context behind why this tool landed when it did. Shopify activated Agentic Storefronts for all eligible US merchants as a default in early 2026. AI traffic to Shopify stores grew 7x and AI-attributed orders grew 11x in the year leading up to that. Most merchants are now connected to ChatGPT, Microsoft Copilot, and Google AI Mode without having made a deliberate choice to be. Being connected and being visible are very different things. The Commerce Readiness Tool is Shopify’s way of showing you the gap.
This article is for merchants who already know agentic commerce is real and want to know what their store actually looks like to an AI agent right now. If you are still building the foundational case for why this matters, the deeper agentic commerce playbook is the right starting point. If you are ready to run the scan and act on what it surfaces, keep reading. We will walk through what the tool tests, what it misses, the three gaps that show up on almost every store, and a stage-aware fix plan that does not assume you have a 12-person dev team.
The Shopify Commerce Readiness Tool runs 31 checks on any public storefront URL, organized into five categories: Agent Discovery, Product Intelligence, Transaction Readiness, Store Quality, and Operational Readiness. No login required, no install, no access to your admin. Paste a product page URL and results come back in about 30 seconds.
One important note on Operational Readiness: it is the only category that is not auto-scanned. It is a short self-check that takes about two minutes and requires no technical knowledge. The other four categories are fully automated against your public storefront.
Five honest limitations worth knowing before you read too much into your score.
First, it does not test live AI recommendations against real prompts. A score of 80 does not mean ChatGPT is recommending your products. It means your store passes 80% of the technical checks that make recommendation more likely.
Second, it does not validate your Google Merchant Center feed quality. The Universal Commerce Protocol that powers Google AI Mode shopping runs largely off your Merchant Center feed data, which lives outside your storefront. The tool cannot see it.
Third, it does not measure brand authority signals. Backlinks, third-party reviews, and external citations all influence AI recommendations. None of them appear in a 30-second storefront scan.
Fourth, it does not check whether AI crawlers are blocked at the network layer. If GPTBot or ClaudeBot are blocked at your WAF or Cloudflare configuration, the tool may not catch it. This is a meaningful gap for stores that have aggressive bot-blocking policies.
Fifth, it does not evaluate the quality of your product titles or descriptions for human shoppers. Structure and language quality are different problems. The tool tests the former, not the latter.
For stores under $500K, this tool is likely the most useful diagnostic you will run all year. For stores at $500K to $5M, it will show you what fell through the cracks of earlier optimization work. For stores above $5M, treat it as a hygiene baseline, not a ceiling.
Running the Commerce Readiness Tool takes about 30 seconds: paste a full product page URL into commerce-readiness.shopify.io, wait for the scan, and read the score and impact-effort matrix that follows.
The tool is designed to scan product pages, not homepages. Product pages contain the richest structured data, the most extractable attributes, and the checks most relevant to AI shopping recommendations. Your homepage will return a thinner result set. Paste the URL of a representative product page, ideally one of your top sellers or a product that represents your core category.
Your overall score is a weighted composite across the five categories. Do not fixate on the number before you look at the category breakdown. A store scoring 55 overall might have Agent Discovery at 30 and Transaction Readiness at 85. Those two stores need completely different fixes. The category breakdown tells you where the weight is concentrated.
This is the most useful output the tool produces. Every failed check is mapped to an impact-effort quadrant: how much will fixing it move your score, and how hard is it to fix. Read this before you do anything else. The matrix is where your 30-day plan lives.
For stores under $500K, scan one product URL to establish your baseline. For $500K to $5M stores, scan 3 to 5 representative product URLs across different categories to determine whether issues are global or product-specific. For stores above $5M, scan a representative sample across product types, then have your dev team run the same scan after every major theme update.
Your overall score is a weighted average across the five categories, but the most useful output is the impact-effort matrix that ranks each failed check by how much it will move your score versus how hard it is to fix.
Shero Commerce published a 1,000-store AI Search Readiness benchmark that gives you a real frame of reference. The average overall AI Search Readiness score across those stores was 42 out of 100. Product pages performed the strongest at an average of 53. Category pages were the weakest at 35. Homepages scored slightly above category pages but still lacked the extractable structure AI agents need.
Vertical differences were significant. Beauty, wellness, and electronics brands scored noticeably higher because their products require detailed explanation and naturally generate richer structured attributes. Lifestyle, apparel, and home goods brands scored lower because they rely more on visuals and lighter copy, leaving AI agents with limited material to summarize.
As a working rule: scoring above 50 puts you ahead of most stores in your category. Scoring 65 to 80 is competitive for mid-market. Scoring 80 or above is the target for stores doing more than $5M annually, with the understanding that marginal gains above 80 may not move actual AI visibility unless you also have strong authority signals.
The decision rule is straightforward. High impact plus low effort means fix it this week. High impact plus high effort means put it on a 30-day plan with a developer. Low impact plus low effort means batch it with weekly maintenance. Low impact plus high effort means ignore it unless you are above $5M and have the team capacity to pursue marginal gains.
Most stores will find that llms.txt (low effort, meaningful signal), product schema completeness (moderate effort, high impact), and policy page rewrites (low to moderate effort, high impact) all cluster in the high-priority quadrant. That pattern holds across categories and revenue bands, which is why those three gaps get their own section below.
The score reflects technical structure. Actual AI visibility is also a function of brand authority, review volume, and category competitiveness. As Adam Finan noted in his breakdown of the Represent audit, these tools are meant to operate as a guide and signal of things to look at, not as a check-the-box exercise. A store with a perfect readiness score and zero third-party reviews still loses to a store with an 80 and 5,000 reviews when an AI agent is deciding which brand to recommend to a shopper who asked for social proof. Use the score as a starting point, not a finish line.
Across stores scanned in the past 30 days, three gaps consistently surface: a missing llms.txt file, incomplete JSON-LD product schema, and thin or missing policy content on returns, shipping, and privacy.
An llms.txt file is a structured plain-text file placed at your domain root (yourdomain.com/llms.txt) that tells AI agents and large language models what your store sells, who you serve, and where to find your most important pages. Think of it as a curated map for crawlers that cannot easily navigate a full Shopify catalog on their own.
The honest position on llms.txt is that there is genuine industry debate about how much weight it carries. Not every AI crawler reads it. Some practitioners argue it is essential. Others, including some who have run live audits on well-performing stores, have questioned whether it meaningfully changes recommendation rates for stores that already have strong structured data. Both positions are defensible with the current evidence.
The practical take for most merchants: for stores doing under $5M, llms.txt is a 10 to 30-minute investment with asymmetric upside. If AI providers expand support, and the trajectory suggests they will, early implementation costs nothing and the potential benefit compounds. For larger stores, prioritize it after product schema is complete. The schema work has more certain impact; llms.txt is the next layer.
For Shopify stores, llms.txt is best generated dynamically from your catalog rather than written manually, because it needs to stay in sync as your products and policies change. Several apps in the Shopify ecosystem can automate this. Manual files go stale quickly on active catalogs.
Structured product data in JSON-LD format is the highest-leverage technical fix available to most Shopify merchants right now. Stores with complete product schema are mentioned in AI shopping responses at roughly 3x the rate of stores with minimal schema. That gap is material.
Most Shopify themes include basic product schema by default, but “basic” typically means name, image, description, and price. What AI agents need to make confident recommendations goes well beyond that. The fields that consistently separate high-performing stores from low-performing ones in AI recommendations are aggregateRating (with actual review count and rating value), shippingDetails, hasMerchantReturnPolicy, and brand. Without those fields, the agent is guessing on the details buyers ask about most.
For a detailed walkthrough of how to structure product data for AI agents, including the specific metafield strategy that drives the most improvement, the guide on how to structure your Shopify product data for AI agents covers the full implementation path.
Apps that help close this gap: Naridon for automated schema enrichment, Judge.me and Yotpo for aggregateRating schema tied to live review data, and Loox for visually-driven brands that need photo review schema. The right choice depends on your review volume and the complexity of your product catalog. For a direct comparison of how reviews drive AI discovery for Shopify brands, the Yotpo review covers the competitive landscape in detail.
For stores under $500K, product schema is the single highest-leverage fix per hour spent. Start here before anything else.
Returns, shipping, and privacy pages that are either missing or written so generically that AI cannot extract specific facts are a consistent gap across every revenue band. This one surprises merchants who assume their policies are fine because they exist.
The problem is not presence. It is specificity. An AI agent checking policies before recommending your store is looking for concrete answers to the questions buyers ask before purchase: how many days for returns, what is the shipping speed to a specific region, what happens if a product arrives damaged, is there a restocking fee. A policy page that says “we offer hassle-free returns” gives the agent nothing to work with. A policy page that says “returns accepted within 30 days of delivery, no restocking fee, refund processed within 5 to 7 business days” gives the agent everything it needs to answer a pre-purchase question and move the buyer toward checkout.
The fix is a rewrite, not a rebuild. Take your existing policy pages and add specific numbers everywhere a vague phrase currently lives. Return window in days. Shipping speeds by region or carrier. Refund timeline. The same answer-first structure that works on product pages works on policy pages. AI agents extract facts from both in the same way. For the underlying data hygiene framework that makes this work across your entire store, the piece on why data hygiene is the single biggest AI commerce lever covers the three-layer approach in full.
At $500K to $5M, fix all three gaps in sequence over 14 days. Schema first, llms.txt second, policy rewrites third. That sequence reflects impact per hour, not arbitrary preference.
The tool diagnoses the technical layer of your store, but it does not measure four things that often determine actual AI visibility: feed quality in Google Merchant Center, content authority and backlinks, review volume and recency, and whether your AI crawlers are reachable at the network level.
Google’s Universal Commerce Protocol, the standard that powers Google AI Mode shopping recommendations, runs largely off your Merchant Center feed data. The Commerce Readiness Tool reads what is publicly visible on your storefront. It cannot see what is in your feed, whether your product categories are mapped correctly to Google’s taxonomy, or whether your feed is suppressing products due to data quality issues. A store can pass every Agent Discovery and Product Intelligence check and still underperform in Google AI Mode because the feed that actually drives those recommendations is broken or incomplete.
If you are above $1M in revenue and Google is a meaningful traffic channel, a feed audit is a separate workstream from the Commerce Readiness scan. Tools like paz.ai can run that audit against your actual feed data rather than your storefront HTML.
AI agents weight third-party reviews, ratings, return rates, and external mentions when deciding which brands to recommend. The tool sees none of this. A store with a perfect readiness score and zero external reviews still loses to a store with an 80 and 5,000 reviews when an agent is evaluating which brand to surface for a buyer who asked for social proof. Authority signals compound over time and cannot be fixed with a 30-day technical sprint. They require a longer-term investment in review generation, press coverage, and community presence that runs parallel to the technical work.
If GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, or Google-Extended are blocked at your WAF, Cloudflare configuration, or robots.txt, the Commerce Readiness Tool may not catch it. This is a meaningful gap for stores that have implemented aggressive bot-blocking policies to manage bandwidth or protect against scraping. Bot policy decisions that made sense six months ago can have real downstream impact on AI crawler access today. Check your robots.txt and your WAF allow-list explicitly. Do not assume the scan would have surfaced a block if one existed.
The tool tests structure, not language. A perfectly schema-marked product titled “BLACK V2 PRO” is still going to underperform a product titled “Lightweight Waterproof Trail Running Jacket, Men’s Black, Size M” in agent queries, because the agent cannot match the first title to a buyer’s natural language request. Schema tells the agent what type of data it is looking at. The title and description tell the agent what the product actually is. Both matter. The tool only evaluates the former.
At all stages, treat the readiness score as one of three or four signals you track, not the only one. At $5M and above, this becomes especially important because protocol-level behavior across UCP, ACP, and MCP is not yet standardized enough to be fully captured in a single automated score.
The right fix plan depends on where your business is: a store under $500K should focus on three quick wins in 30 days, a $500K to $5M store should run a 60-day systematic fix across all five categories, and a store above $5M should treat the score as a starting baseline for a deeper protocol-level audit.
Day 1: run the scan on your top-selling product URL. Screenshot the score, the category breakdown, and the impact-effort matrix before you change anything. You need a baseline to measure against.
Days 2 to 5: generate your llms.txt file. Use an app rather than writing it manually so it stays in sync with your catalog. This is a 10 to 30-minute setup with no ongoing maintenance cost if you automate it.
Days 6 to 14: complete your product schema. Use Naridon or a similar app to automate schema enrichment across your catalog rather than editing product templates manually. Prioritize aggregateRating (requires an active review app like Judge.me or Loox), shippingDetails, and hasMerchantReturnPolicy. These are the fields most commonly missing from basic Shopify theme schema.
Days 15 to 21: rewrite your policy pages with specific numbers. Return window, shipping speeds, refund timeline. Thirty minutes per policy page is enough to add the specificity AI agents need.
Days 22 to 30: re-scan using the same product URL and document the score lift. Attribute the change to the specific fixes you made. That documentation becomes your baseline for the next improvement cycle.
Weeks 1 to 2: complete everything in the 30-day plan above, but scan 3 to 5 product URLs across your top categories to identify whether issues are global or product-specific. Global issues get fixed at the theme level. Product-specific issues get fixed at the catalog level.
Weeks 3 to 4: optimize your top 20 products by traffic for FAQ schema. Add structured Q&A blocks to product pages and implement FAQ schema markup. This is the single most underused structured data type in the Shopify ecosystem and one of the strongest signals for AI recommendation eligibility.
Weeks 5 to 6: run a product description audit on your top 10 SKUs by revenue. Rewrite descriptions to lead with machine-parseable facts: material, dimensions, weight, certifications, compatibility, use case tags. Brand story and emotional copy can follow, but the structured facts need to come first.
Week 7 to 8: re-scan across all product URLs from your initial sample, document category-level score changes, and set a quarterly re-scan cadence going forward.
Run the Commerce Readiness scan as a hygiene check, not a comprehensive audit. It will surface obvious gaps, but your real work is at the protocol level. Layer in a Google Merchant Center feed audit using paz.ai or a comparable tool. Run a backlink and citation review to understand your brand authority signals relative to category competitors. Audit your product taxonomy against Google’s Product Taxonomy standard to ensure your feed categories are mapped correctly for AI Mode recommendations. Set a quarterly re-scan cadence and have your dev team run the scan after every major theme update or app change.
At this stage, the in-house dev team or an external partner should be doing protocol-level work that goes well beyond 31 checks. The Commerce Readiness Tool is the floor of your audit process, not the ceiling.
The Commerce Readiness Tool launched roughly a month after Shopify activated Agentic Storefronts for every eligible US merchant by default in early 2026, which means most stores are now connected to ChatGPT, Microsoft Copilot, and Google AI Mode whether they realized it or not.
In the Shopify Agentic Storefronts announcement, Tobi Lutke described the goal as making every Shopify store agent-ready by default. That language is precise. Agent-ready by default means the infrastructure is on. It does not mean your store is optimized, visible, or competitive within that infrastructure. Shopify reported a 15x increase in AI-originated orders over the past year. AI traffic to Shopify stores grew 7x in the same window. Those numbers represent a channel that is already live and already influencing revenue for merchants who are ready for it.
The Commerce Readiness Tool is Shopify’s way of telling merchants: we put you in the room. Here is how to make sure they can actually see you. Shopify auto-enrolled millions of merchants into Agentic Storefronts but did not auto-fix their data. The scan is the diagnostic layer that was always going to be needed once the infrastructure was live. The fact that it is free, requires no login, and returns results in 30 seconds is a deliberate choice to lower the barrier to entry for merchants at every revenue band.
For the deeper strategic context on what agentic commerce means for Shopify merchants, including the protocol landscape and the seven-step readiness framework, the deeper agentic commerce playbook covers the full picture. This article is focused on the tool itself and what to do with your score. Those two pieces are designed to be read together, not as substitutes for each other.
Stores above $5M tend to assume their existing tech stack handles AI readiness. That assumption is usually wrong. The Commerce Readiness scan is worth running regardless of how sophisticated your existing optimization work is, because the categories it tests are specific to agentic commerce and distinct from traditional SEO or conversion optimization.
The Commerce Readiness Tool gives you a ranked list of priorities. The only thing that produces an actual score lift is execution against that list within 30 to 60 days, before the scoring criteria evolve.
The scan takes 30 seconds. Reading and understanding the output takes another 10 to 15 minutes. The remaining time in your first two-hour session should go toward identifying which items in the impact-effort matrix fall into the high-impact, low-effort quadrant and assigning them to whoever in your organization owns technical implementation. For most stores under $1M, that is a developer or a technically capable founder. For stores above $1M, it is typically a dev team with a product manager coordinating.
Most merchants fix things and then cannot remember what their pre-fix score was. Take screenshots of three things before you touch anything: the overall score, the category breakdown, and the full impact-effort matrix. Store them somewhere your team can reference. When you re-scan in 30 days, you want to be able to attribute specific score changes to specific fixes. That attribution is what turns a one-time scan into a repeatable improvement process.
If you have a developer and your score is above 40, the priority list from the scan is enough to direct their work. If you scored under 40, if your store is above $1M and you do not have the internal capacity to systematically work through the list, or if the impact-effort matrix surfaces issues that require judgment beyond technical implementation, that is the case for outside help. The technical fixes are straightforward. The judgment layer, knowing which gaps actually matter for your category and your competitive set, is where experience with AI visibility across many stores makes the difference.
[CTA PLACEHOLDER: Steve to finalize the exact wording and link target for the AI Visibility services CTA. Suggested framing: “If you ran the scan and the priority list looks bigger than your team can absorb, this is exactly the kind of work the eCommerce Fastlane AI Visibility audit is built to translate into a 30-day execution plan.”]
It is a free, no-login scanner that Shopify launched in late April 2026 that runs 31 checks on any public storefront URL across five categories: Agent Discovery, Product Intelligence, Transaction Readiness, Store Quality, and Operational Readiness. There is no cost, no install, and no Shopify account required. You paste the URL of any product page on any public Shopify store and the scan returns in about 30 seconds. The result is a score, a category breakdown, and an impact-effort matrix that ranks the gaps by what will move your score the most for the least development time. The Operational Readiness category is the one exception to the auto-scan: it is a short self-check you complete manually, which takes about two minutes and requires no technical knowledge.
There is no official benchmark from Shopify, but third-party data from a benchmark of 1,000 Shopify stores put the average AI Search Readiness score at 42 out of 100, with product pages averaging 53 and category pages averaging 35. As a working rule, scoring above 50 puts you ahead of most stores in your category, scoring 65 to 80 is competitive for mid-market stores, and scoring 80 or above is the target for stores doing more than $5M annually. Beauty, wellness, and electronics stores tend to score higher than apparel or home goods because their product pages contain more extractable detail by nature of the category. A score above 80 does not guarantee AI visibility if your authority signals, review volume, and feed quality are not also strong.
The Shopify Commerce Readiness Tool is built and operated by Shopify and runs 31 checks specifically calibrated to what Shopify’s own infrastructure (Agentic Storefronts, the Shopify Catalog) and partner AI agents (ChatGPT, Microsoft Copilot, Google AI Mode, Perplexity) actually evaluate. Third-party scanners from companies like FoundGPT, Naridon, and VisionTags use their own scoring methodologies, often with different check sets and weightings. The Shopify tool is best used as the starting baseline because it is free, fast, and calibrated to Shopify’s own platform defaults. Third-party tools are best used as the next layer once you have acted on the Shopify scan results, particularly for feed-level audits and competitive benchmarking that the Shopify tool does not cover.
An llms.txt file at your domain root tells AI agents and large language models what your store sells, who you serve, and where to find your most important pages, in a structured format they can parse without navigating your full catalog. The honest answer is that not every AI crawler reads llms.txt yet, and there is genuine debate in the practitioner community about how much weight it carries relative to structured data and feed quality. But it is a 10 to 30-minute one-time setup with no meaningful downside and real upside if AI providers expand support, which the current trajectory suggests they will. For Shopify stores, llms.txt is best generated dynamically from your catalog using an app rather than written manually, because it has to stay in sync as your products, policies, and top pages change over time.
Run it once now to establish your baseline, then re-scan after every major change to your store: theme update, app addition or removal, schema implementation, policy page rewrite, or product taxonomy change. As an ongoing cadence, monthly is appropriate for stores under $1M, every two weeks for $1M to $10M, and weekly for stores above $10M or for any store actively running an AI visibility improvement project. Document your score after each scan so you can attribute lift to specific changes rather than guessing what moved the needle. The scan takes 30 seconds, so the cost of running it frequently is essentially zero. The cost of not running it is missing drift that accumulates silently after theme or app updates.
