How to prep your Shopify catalog for agentic AI shopping — the complete 2026 playbook.

The agents.md, llms.txt, and llms-full.txt discovery files are the protocol layer that tells AI crawlers exactly what your store is, what it sells, where the sitemap lives, what your policies are, and how to reach your catalog endpoints. Shopify generates default versions of all three automatically when you enable Agentic Storefronts, but customizing them gives you control over the context AI agents receive before they ever query your product data.

The setup path: in your Shopify admin, go to Online Store › Themes, click Edit code on your active theme, then Add a new template and select agents.md, llms.txt, or llms-full.txt. All three URLs should return consistent content — Shopify’s documentation is explicit that agents.md is the canonical source of truth, while llms.txt and llms-full.txt serve older crawlers. By default they are identical, which is correct. Custom edits are most valuable for multi-brand stores or stores with unusual catalog structures where the default context is insufficient. This work complements broader Shopify AI SEO investments.

Stormy AI is purpose-built for the agentic commerce preparation workflow. It scans your Shopify catalog for the exact issues that cause AI agents to deprioritize your products: missing metadata fields, broken variant configurations, inconsistent attribute values across the same product family, and inventory data that cannot be resolved in real time. Their 2026 guide documented that merchants spend 40% of their implementation time on legacy data cleanup, and Stormy AI exists to compress that timeline.

It also integrates with Shopify Sidekick Pulse to monitor demand spikes before they appear in your sales reports, letting you set proactive fulfillment alerts. The limitation is that Stormy AI is primarily a diagnostic and monitoring tool — it surfaces the issues clearly, but the actual remediation requires you to act on the findings or connect another tool to do it. For stores that want a scanner as part of a broader stack, it earns its place. For stores that want a single autonomous system that finds and fixes catalog issues continuously, Ryze AI covers both sides of the equation.

Shopify Flow is the native automation engine that connects operational triggers (inventory drops below threshold, a product sells out, a price changes) to downstream actions that keep your Shopify Catalog data accurate in real time. This matters enormously for agentic commerce: an AI agent that queries your catalog and receives stale inventory or incorrect pricing data will either recommend a product the shopper cannot buy, or worse, initiate a checkout that fails at fulfillment. Both outcomes damage the customer relationship and reduce your visibility score with the AI platform.

The most impactful Flow workflows for agentic readiness: automatic out-of-stock tagging that agents can read, price change notifications that trigger catalog re-indexing, and low-inventory alerts that prevent agents from promoting products you cannot fulfill. Fabric’s 2026 guide to agentic commerce on Shopify also recommends Flow as the “Prompt-to-Flow architecture” layer — dynamic data triggers that external agents can query via API. Combine this with automated marketing workflows to keep both agents and human shoppers seeing accurate information.

fabric NEON extends Shopify’s product catalog with real-time visibility across all channels and third-party inventory sources, addressing one of the hardest agentic commerce problems for large retailers: when an AI agent asks “can this item be delivered to a customer’s address by a specific date?”, the answer requires querying multiple warehouse locations, carrier APIs, and fulfillment SLAs simultaneously. NEON provides that unified inventory layer on top of your Shopify store.

For enterprise Shopify Plus merchants with BOPIS requirements, ship-from-store workflows, or multi-location order routing, NEON adds the composable fulfillment intelligence layer that Shopify Functions alone cannot provide. fabric’s 2026 agentic commerce guide positions NEON as the bridge between Shopify’s catalog data and the complex fulfillment questions agents need to resolve before completing a purchase. Most stores under 500 SKUs and single-warehouse operations will not need this level of infrastructure — Shopify Flow plus clean Catalog Mapping achieves comparable results for simpler fulfillment models.

The manual catalog audit using Shopify’s built-in bulk metafield editor and CSV export/import is the starting point most merchants land on before discovering that it does not scale. The process is conceptually simple: export your catalog to a spreadsheet, audit every product against Shopify’s recommended field list (title, description, images, product type, vendor, collections, tags, barcode/GTIN/UPC, variants including option names, and all custom metafields for attributes like material, dimensions, sizing, care instructions, and comparison data), fill in every blank, and re-import.

For a 50-SKU store doing a one-time cleanup before enabling Agentic Storefronts, this is perfectly adequate. For a 2,000-SKU store with seasonal launches and frequent price changes, it is a recurring burden that will always be partially out of date by the time you finish. The deeper problem is that manual audits have no monitoring layer — you identify today’s gaps but have no system that flags tomorrow’s. Combining this initial audit with an automated tool like Ryze AI for ongoing catalog health monitoring is the practical path for any store that expects to grow. See also our overview of agentic AI readiness for Shopify for the full checklist.