# AI Agent Readiness Score — methodology and checklist Free tool: https://starlix.net/ai-readiness · Methodology version 2 · Updated 2026-10-29 · Publisher: StarlixAI ## What is AI agent readiness? AI agent readiness is how well an AI shopping assistant such as ChatGPT, Gemini or Perplexity can answer a shopper's questions about your store using only what it can read: your product pages, policies and structured data. A ready store states fit, materials, delivery time, returns and price in plain text and in machine-readable form, so the agent quotes facts instead of guessing or recommending another store. ## How the score works The score is out of 100 and uses 15 fixed rules on public data (no AI), so the same store always gets the same score. The tool reads up to 10 products, spread across the first 30 in /products.json, each product page's schema.org JSON-LD, the shipping, refund, contact information and terms of service policies, robots.txt, sitemap.xml, and the about and contact pages. The score is Starlix's estimate from public pages. It is not a Shopify, OpenAI or Google score, and no score guarantees that an AI agent will recommend a store. ## Product facts agents can quote (30 points) The details a shopper needs to choose: fit, materials, type, brand and a product identifier. ### Size, fit or specs (12 points) - Check: Product descriptions mention a size chart, measurements, dimensions, fit notes or compatibility. - Fix: Add a size chart or measurements to every product, plus one fit note such as "runs small, size up". ### Materials or ingredients (8 points) - Check: Product descriptions say what the product is made of. - Fix: State the material or ingredients with percentages, and add care instructions. ### Product type, brand and option names (5 points) - Check: Each product has a product type and a vendor, and its options have real names like Size or Color. - Fix: Fill in Product type and Vendor in Shopify admin, and rename options from "Title" to Size, Color or Style. ### Brand and GTIN or SKU (5 points) - Check: The product page's structured data includes a brand and a GTIN, SKU or MPN. - Fix: Add barcodes (GTIN) and SKUs to your variants. Most Shopify themes then publish them in structured data. ## Policies agents can quote (25 points) Delivery time and cost, the return window and conditions, and who runs the store. ### Shipping time and cost (10 points) - Check: A public shipping policy exists (4 points), states delivery times (4) and states cost or free shipping (2). - Fix: Write a shipping policy in Settings > Policies with delivery days per region and the cost or free-shipping threshold. ### Return window and conditions (10 points) - Check: A public refund policy exists (4 points), states a return window in days (4) and states conditions (2). - Fix: State the return window in days, the item condition required, and who pays for return shipping. ### Contact information and terms (5 points) - Check: Shopify's contact information policy (3 points) and terms of service (2) are published. - Fix: Publish the contact information and terms of service policies in Settings > Policies. ## Machine-readable data (20 points) Structured data, a product sitemap and robots.txt rules that let agents read your catalog reliably. ### Product structured data (10 points) - Check: Product pages publish schema.org Product JSON-LD with price, currency and availability. - Fix: Use a theme that outputs Product JSON-LD, or an SEO app that adds it. Test a product page with Google's Rich Results Test. ### AI crawler access (5 points) - Check: robots.txt lets GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot and Google-Extended read product pages (1 point each). - Fix: Review robots.txt.liquid. Blocking AI crawlers is a valid choice, but it can keep your products out of those assistants. ### Product sitemap (5 points) - Check: sitemap.xml exists and lists products (5 points), or exists without products (2). - Fix: Shopify generates sitemap.xml automatically. If products are missing, check that they're published to the Online Store. ## Offer clarity (15 points) Variants, stock, prices and images that an agent can describe without guessing. ### Named variants in stock (8 points) - Check: Variants have real names (not "Default Title") and at least one is in stock. - Fix: Name every variant and hide or mark sold-out products so agents don't recommend what you can't ship. ### Consistent sale prices (4 points) - Check: Any compare-at price is higher than the selling price. Stores without compare-at prices get full points. - Fix: Clear compare-at prices that are equal to or lower than the price, so agents don't report a fake discount. ### Product images (3 points) - Check: Each product has at least three images. - Fix: Add at least three images per product: front, back or detail, and in use. ## Trust signals (10 points) Reviews and pages that show a real business stands behind the products. ### Product reviews (6 points) - Check: Product pages publish a rating in structured data or show a review widget. - Fix: Collect product reviews with a reviews app that also publishes ratings in structured data. ### About and contact pages (4 points) - Check: An about page (2 points) and a contact page (2) exist. - Fix: Add an About page with who you are and where you ship from, and a Contact page with a reply time. ## Score bands - 80–100, Agent-ready: AI agents can answer most shopper questions from your store data. - 60–79, Partly ready: AI agents can find your products but will guess on key questions. - 40–59, At risk: AI agents are likely to skip or misdescribe several products. - 0–39, Not ready: AI agents have too little to work with. ## Sources - [schema.org Product](https://schema.org/Product): The vocabulary behind the structured data, brand and identifier checks. - [Google Search Central: merchant listing structured data](https://developers.google.com/search/docs/appearance/structured-data/merchant-listing): Price, currency, availability, GTIN and aggregateRating as machine-readable product facts. - [OpenAI: overview of OpenAI crawlers](https://platform.openai.com/docs/bots): The GPTBot and OAI-SearchBot user agents in the crawler check. - [Anthropic: how site owners can manage ClaudeBot](https://support.claude.com/en/articles/8896518-does-anthropic-crawl-data-from-the-web-and-how-can-site-owners-block-the-crawler): The ClaudeBot user agent in the crawler check. - [Perplexity: Perplexity crawlers](https://docs.perplexity.ai/guides/bots): The PerplexityBot user agent in the crawler check. - [Google: common crawlers, including Google-Extended](https://developers.google.com/search/docs/crawling-indexing/google-common-crawlers): The Google-Extended token in the crawler check. - [Shopify Editions Fall '26: agentic commerce](https://www.shopify.com/editions/fall2026): Shopping through AI agents, and compare-at prices shown in AI shopping agents. - [Shopify Knowledge Base app](https://apps.shopify.com/shopify-knowledge-base): AI agents answering from a store's policies, settings and FAQs. ## Limits - No AI platform publishes how it chooses or ranks stores. The checks are signals those platforms document or that shoppers ask about, not a ranking formula. - The point weights are Starlix's judgment, weighted toward the questions shoppers ask most before buying (fit, delivery, returns). We will adjust them as we compare scores with real agent answers. - Only public pages are read, and only up to 10 products. Large catalogs, headless storefronts and policies hosted on other domains can score lower than they should. - Text checks use keyword rules, so a size chart shown only as an image, or written in another language, may not be detected. ## FAQ ### What is AI agent readiness? AI agent readiness is how well an AI shopping assistant such as ChatGPT, Gemini or Perplexity can answer a shopper's questions about your store using only what it can read: your product pages, policies and structured data. A ready store states fit, materials, delivery time, returns and price in plain text and in machine-readable form, so the agent quotes facts instead of guessing or recommending another store. ### What do AI shopping agents read about my store? Mostly public data: product titles and descriptions, prices, variants and stock, your shipping and refund policies, reviews, and the schema.org structured data on product pages. Some agents also use catalog feeds and knowledge bases provided by the platform. The clearer and more complete that data is, the fewer questions an agent has to guess. ### How is the score calculated? With fixed rules, not AI, so the same store gets the same score. We run 15 checks in five groups: product facts (30 points), policies (25), machine-readable data (20), offer clarity (15) and trust signals (10). The full checklist with points and fixes is on this page and in a plain-text version at starlix.net/ai-readiness.md. ### How do I improve my AI agent readiness score? Start with the biggest gaps in your report. For most stores that means adding a size chart or measurements to every product, stating delivery days and the return window in your policies, and making sure product pages publish structured data with price and availability. Each gap in the report includes the specific fix. ### Is this an official Shopify, OpenAI or Google score? No. It's Starlix's estimate from your public pages. No AI platform publishes how it ranks stores, and a high score doesn't guarantee that any agent will recommend you. It shows which shopper questions your store data can and can't answer. ### How is this different from SEO? SEO is about ranking a page in search results. Agent readiness is about whether an assistant can answer a specific question from your data, such as fit, delivery time or the return window. Good SEO helps, but a page can rank well and still leave those questions unanswered. ### Does Shopify's Knowledge Base app replace this? They work together. Shopify's free Knowledge Base app lets you review and customize the FAQs AI agents use for your store. This check looks at the product pages, policies and structured data those answers are built from, and shows what's missing. ### Should I let AI crawlers into my store? That's your call. We only report what your robots.txt allows. Blocking a crawler can keep your pages out of that assistant's answers, and some stores block crawlers on purpose. ### Do I need to install anything or log in? No. The check only reads pages any visitor can see. It doesn't need access to your Shopify admin. ### Why can't it scan my store? The store may be password-protected, may not be on Shopify, may hide its product catalog, or may ask scanners not to read it in robots.txt. We respect all of these.