The diagnostic flow
The eligible-but-invisible problem has five common causes, in descending order of frequency and ascending order of cost-to-fix. Catalog eligibility (mechanical, fix in minutes). Per-channel access settings (admin toggle, fix in minutes). robots.txt.liquid AI-bot blocks (theme code edit, fix in 15 minutes). Product-data quality (PDP rewrites, fix in days). Category capture (citation density, fix in months). Walk them in that order; you'll find the cause before you get to the hardest gate.
1. Catalog eligibility
Open Shopify admin. Confirm the store is on the Starter plan or higher, not password-protected, and ships to the United States or Canada. Open three products. Confirm each has a title, at least one image, a price over $0, an identifiable URL, and is not set to Unlisted status. If any one of those eleven gates fails, the product is out of Catalog — and out of every AI surface that reads it.
2. Channel access
Open admin.shopify.com/agentic. Review which AI channels are enabled. Confirm the products you expect to surface in (say) ChatGPT or Perplexity have the Shopify Catalog access toggle enabled for that channel and the Direct checkout toggle set the way you want. Merchants who silently disabled a channel in setup forget about that decision six months later — this is where it shows up.
3. robots.txt.liquid audit
Open the theme code editor. Look for templates/robots.txt.liquid. If the file is absent, Shopify defaults apply — the store allows AI crawlers, and this gate passes. If the file is present, audit it for Disallow rules under any of these user-agents: GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-User, Claude-SearchBot, PerplexityBot, Google-Extended. Pasted-in 2024 'block AI bots' templates are the most common cause of eligible-but-invisible.
4. Product-data audit
Open three products. Score each against Shopify's seven AI-input categories: detailed specifications, comparison information, comprehensive descriptions, structured data, sizing/material/care, price/availability/key features, and alt text on every image. A product missing any three of these is unlikely to surface in AI recommendations regardless of Catalog eligibility — there isn't enough structured signal for the AI to disambiguate.
5. Category capture
If gates 1-4 all pass and the products still don't surface, the answer is usually category capture. In some categories, two or three brands hold 80%+ of AI citations. GLP-1 medications: two pharma companies own nearly 100% of citations across ChatGPT, Claude, and Perplexity. Facial skincare: La Roche-Posay holds 81% share. A new entrant in these categories isn't competing on product data — it's competing on editorial citation density on third-party sites, which AI engines train on.
The 4 articles under this cluster
Four articles cover the diagnostic in depth. The 'not showing up' article walks the long-form flowchart with every step. The category-capture article covers gate five specifically. The AI visibility check tool article describes the criteria a real check should cover. The 30-day report article shows what an audit report should output.