CatalogScan

AI Readiness

Ecommerce AI Readiness Checklist 2026: 4 Tiers for AI Shopping Agents

ChatGPT Shopping, Perplexity Commerce, and Google AI Mode have replaced a meaningful share of top-of-funnel product search. Getting your store into their recommendations requires passing a set of technical signals most merchants don't know exist. This checklist prioritizes those signals by impact and time to fix — from 30-minute critical fixes to multi-week catalog projects.

TL;DR Work through the tiers in order. Tier 1 (30 min) blocks everything if broken — fix it first. Tier 2 (1-2 days) covers the structured data quality signals AI agents score products on. Tier 3 (1-2 weeks) adds review signals and variant linking that moves you ahead of 60% of competitors. Tier 4 (ongoing) is the advanced work that separates the top 10% of AI-visible stores.

Tier 1 — Critical access gates (≈30 minutes)

These items block AI indexing entirely if they fail. Fix these before touching anything else — optimizing structured data on a crawl-blocked store is wasted effort.

  • Product feed accessible at /products.json Test: curl -sI -A "GPTBot/1.0" https://yourdomain.com/products.json. Expect HTTP 200 and Content-Type: application/json. A 403, 401, or HTML challenge page means AI agents cannot bulk-ingest your catalog. Fix: review Cloudflare Bot Fight Mode, Shopify Liquid robots.txt, and any IP-restriction apps. See product feed URL guide.
  • AI crawlers not blocked in robots.txt Test: curl -s https://yourdomain.com/robots.txt. Look for User-agent: GPTBot, OAI-SearchBot, ClaudeBot, or PerplexityBot followed by Disallow: /. Also check for a blanket User-agent: * / Disallow: /. Fix: edit your Shopify theme's robots.txt.liquid to explicitly allow AI crawler user-agents.
  • Cloudflare Bot Fight Mode not blocking AI crawlers Test with curl -sA "GPTBot/1.0" https://yourdomain.com/products/[any-product]. If the response is "Just a moment... Cloudflare" you have a WAF/Bot Fight Mode issue. Fix: Cloudflare dashboard → Security → Bots → disable Bot Fight Mode or add AI crawler IPs/user-agents to the allowlist. See Cloudflare AI crawlers guide.
  • Store returns HTTP 200 (not 503, redirect loop, or auth gate) Test: curl -sI https://yourdomain.com. A password-protected store (Shopify "store coming soon" mode) returns 401 to crawlers. A misconfigured redirect chain returns 301/302 in a loop. Both block indexing. Fix: disable Shopify's "Password protect this store" setting under Online Store → Preferences.

Tier 2 — Structured data quality (1–2 days)

Once access gates pass, AI agents read your structured data to evaluate product quality and match against buyer queries. These items determine whether your products are cited in recommendations or skipped in favor of better-documented competitors.

  • Product JSON-LD on every product page Test: view source on a product page and search for application/ld+json with "@type": "Product". Most Shopify themes include this, but headless storefronts (Hydrogen, Next.js) often omit it after migration. Required minimum fields: name, description, image, offers (with price, priceCurrency, availability). See schema markup guide.
  • GTIN populated on at least 80% of variants Test: curl -s https://yourdomain.com/products.json | python3 -c "import json,sys; p=json.load(sys.stdin)['products']; vars=[v for pr in p for v in pr['variants']]; covered=sum(1 for v in vars if v.get('barcode')); print(f'{covered}/{len(vars)} ({100*covered//len(vars)}%)')". GTINs enable exact-match variant queries in ChatGPT Shopping and price-comparison indexing in Google Shopping Graph. Fix: bulk import from supplier sheets. See GTIN guide.
  • Product descriptions 150+ words on key products AI agents use description text to match long-tail queries. Descriptions under 100 words lack enough signal density for citation. Priority: your top 20% of SKUs by revenue. Use the AI description generator guide for prompt templates that produce signal-rich copy.
  • Brand field in Product JSON-LD is an entity, not a string Correct format: "brand": {"@type": "Brand", "name": "YourBrand"}. Most themes output a bare string — this reduces entity resolution confidence in AI systems that maintain a product knowledge graph. Fix in theme's product.json or JSON-LD Liquid block.
  • Canonical URL on every product page (matching feed URL) Shopify generates canonical tags automatically. Verify: view source → search for rel="canonical". The canonical should point to /products/[handle], not a collection URL. Also verify your Merchant Center feed uses the same URL format — mismatches fragment your Shopping Graph entry.

Tier 3 — Trust and citation signals (1–2 weeks)

Tier 3 signals separate products that AI agents reference from products they actively recommend. These signals tell the AI system whether your product is trustworthy and how it relates to other products in the same family.

  • AggregateRating in Product JSON-LD (review stars) The highest-impact Tier 3 fix. Perplexity Commerce uses aggregateRating as a recommendation confidence signal — products with reviews are cited 3× more often than identical products without. Enable via your review app (Judge.me, Yotpo, Okendo, Stamped.io) — each has a "Rich snippets" or "Schema.org" toggle in settings. See AggregateRating guide.
  • ProductGroup JSON-LD linking variants to parent product Without ProductGroup, a Shopify product with 20 color/size variants appears as 20 separate products in AI shopping panels — each with a lower signal-to-noise ratio. With ProductGroup, they consolidate into one card with variant selection. 63% of Shopify stores emit ProductGroup incorrectly or not at all. See ProductGroup JSON-LD guide.
  • FAQPage JSON-LD on product pages or category pages Perplexity Citations appear at the bottom of AI answers — FAQ schema on your product page increases the chance Perplexity uses your page as a source when answering buying questions about your product category. Minimum 3 Q&A pairs addressing "how does this compare to [alternative]," "what is the return policy," and "what materials are used." See Perplexity Commerce optimization guide.
  • Sitemap includes all product and collection pages Verify: curl -s https://yourdomain.com/sitemap.xml should reference a sitemap_products_1.xml sub-sitemap. Headless storefronts often generate incomplete sitemaps. Also verify no products are excluded via noindex meta tag that you did not intentionally exclude.
  • Google Merchant Center feed approved (for Google AI Mode) Google AI Mode pulls recommendations from the Shopping Graph, which requires an approved Merchant Center feed. Without a feed, your store may appear in AI Mode text but not in product card panels (which show real-time price and stock). Setup: Google & YouTube Shopify app → Merchant Center. See Google Shopping feed setup guide.

Tier 4 — Advanced visibility (ongoing)

Tier 4 items move you ahead of competitors who have already done Tiers 1–3. These require more investment but compound over time.

  • Shopify Standard Product Taxonomy metafields populated Shopify's Standard Product Taxonomy metafields (custom.shopify.*) map to Google Shopping attributes automatically when the Google & YouTube app syncs. Filling in material, age_group, target_gender, and product_type improves category classification in the Shopping Graph. See Shopify metafields AI agents guide.
  • Perplexity Merchant Verified program (if eligible) Perplexity's Merchant Verified program gives verified stores a trust badge in Commerce results. Application at perplexity.ai/merchant. Eligibility: active domain, Merchant Center (or equivalent), and product feed accessible. Verified status increases citation frequency in Perplexity Commerce answers. See Perplexity Commerce guide.
  • Description richness across the full catalog (150+ words) Tier 2 targeted your top 20% of SKUs. Tier 4 extends this to the full catalog. Use batch AI generation (ChatGPT/Claude API) with your product export CSV. See the AI product description generator guide for prompt templates and bulk update workflow.
  • MPN (Manufacturer Part Number) populated for resale products For resellers carrying third-party brands, the MPN is the secondary cross-retailer identifier (after GTIN). AI agents use MPN to consolidate your listing with the same product at competing retailers in price-comparison queries. Add via mm-google-shopping.mpn metafield. Check your supplier's product sheet — MPN is typically in the SKU or product spec column.
  • Monitor and maintain: re-scan monthly AI agent signals drift. Common regression causes: Shopify theme updates changing JSON-LD structure, new app installs adding duplicate schema blocks, Cloudflare rule changes re-blocking crawlers, new product variants added without GTIN. Schedule a monthly CatalogScan to catch regressions before they affect recommendation volume.

Common questions

Do I need to complete all 4 tiers to appear in AI recommendations?

No. Tier 1 is mandatory — a blocked store appears in nothing. Tiers 2 and 3 are where most of the recommendation volume comes from. Tier 4 is competitive differentiation. A store that completes Tier 1 and Tier 2 is already ahead of 60% of Shopify merchants based on our 100-store scan data. The value of Tiers 3 and 4 is compounding: each signal you add slightly increases your recommendation frequency, and the effects stack.

Is this checklist specific to Shopify or does it apply to all ecommerce platforms?

The signals are platform-agnostic — ChatGPT Shopping, Perplexity, and Google AI Mode don't care what platform you're on. The Tier 1 details (products.json URL, robots.txt.liquid) are Shopify-specific. For WooCommerce, Magento, or BigCommerce, the equivalents are different product feed paths and theme template files, but the structured data requirements (Product JSON-LD, GTIN, AggregateRating, ProductGroup) are identical. The CatalogScan audit works on any platform that serves a product page over HTTP.

My store is headless (Hydrogen / Next.js). Which items are most affected?

Headless stores are at higher risk on Tier 1 (products.json accessibility if not proxied) and Tier 2 (Product JSON-LD often missing after migration from Liquid themes). Tier 1 Cloudflare item may also fail more frequently because headless architectures often add a CDN/WAF layer. Check all four Tier 1 items carefully. For JSON-LD in headless, add a structured data component to your product page template that pulls data from the Storefront API — most Next.js/Hydrogen templates have a slot for this in the document head.

How do I know which tier I'm currently at?

Run a CatalogScan — it audits 18 signals and maps them to approximate tiers, showing which are passing and which are failing. Alternatively, run the four curl tests in Tier 1 manually (they take under 10 minutes), then use Google Rich Results Test on a product page for Tier 2. If Tier 1 passes and Rich Results Test shows your Product JSON-LD with no errors, you're working on Tier 3.

Does completing this checklist guarantee AI shopping placement?

No tool or checklist can guarantee placement — AI recommendation systems make probabilistic decisions based on query matching, competitive landscape, and real-time inventory. What the checklist guarantees is eligibility: a store that fails Tier 1 is ineligible by definition. A store that completes Tiers 1–3 is eligible and competitive. Placement frequency then depends on query volume for your products, competitor signal quality, and review score relative to alternatives.

Check where your store stands across all 4 tiers — CatalogScan audits 18 signals and shows your top 5 fixes prioritized by impact.

Run the free CatalogScan →

Related reading