The Agentic Commerce Readiness Checklist

Every merchant asks some version of the same question before they call anyone: is my store ready for AI shopping agents? You don't need a consultant to find out. Most of what determines readiness is visible from your own admin, your own product pages, and ten minutes with a shopping assistant. This is the checklist we run internally before scoping any Agentic Commerce Setup engagement — now in a form you can run yourself first.

Section 1 — Catalog attribute coverage

This is where readiness is won or lost. Agents — Shopify's Brand Agents, Amazon's Rufus, and off-site assistants like ChatGPT and Claude — reason over structured fields, not marketing paragraphs. If a fact about your product lives only inside a description, an agent can't reliably use it to answer a shopper's question. Full background on why this matters is in the catalog enrichment playbook; here's the checklist version.

1. 1

   Pull your top 50 SKUs by revenue and list every attribute a shopper would ask about

   Material, size, compatibility, care instructions, certifications — whatever is category-specific. Write the list before you look at your data, so you're not anchored to what already exists.

2. 2

   Check what percentage of those fields are actually populated

   Most mid-market catalogs land between 40% and 60% coverage on the attributes that matter. Anything under 80% on your hero SKUs is a gap worth fixing before anything else on this list.

3. 3

   Search your product descriptions for facts that never became fields

   Dimensions mentioned in a sentence, materials named in passing, compatibility claims buried in copy. These are prose-trapped specs — the fact exists, but no agent can extract it reliably.

4. 4

   Check for compatibility and "works with" data

   This is the most commonly missing attribute type and the highest-value one. Shoppers ask agents "does this fit a Weber Spirit II" or "will this work with my iPhone 16 Pro" constantly. If you have no compatibility relationships in your data, the agent has nothing to match against.

Section 2 — Structured data an agent can actually parse

Attribute coverage is about what's in your PIM or metafields. Structured data is about whether that information is exposed on the page in a format machines parse — Product, Offer, and Review schema, at minimum. This is a separate check from catalog completeness: you can have rich attributes in Shopify metafields that never make it into the page's schema markup at all.

• Product schema present on every PDP — name, image, description, SKU, brand — not just on a handful of flagship pages.
• Offer schema with live price and availability — stale or missing availability is one of the fastest ways to lose agent trust, since a wrong in-stock claim breaks a transaction downstream.
• AggregateRating and Review schema, if you collect reviews — agents weight review evidence heavily; see reviews as machine-readable evidence for the Amazon-side version of the same principle.
• Breadcrumb and Organization schema for category context and brand identity resolution.

For the full audit method — including how to test what's actually being read versus what's merely present in your theme's source — see structured data for AI shopping and the AI citation audit guide.

Section 3 — PDP server-rendering

How do I know if my catalog is agentic-commerce ready? Start here, because it's the fastest yes/no test: fetch one of your product pages the way an agent's crawler does — a plain HTTP request, no JavaScript execution — and check whether the price, availability, and core attributes are present in the raw HTML.

1. 1

   Disable JavaScript and reload a PDP

   Use your browser's dev tools to simulate a no-JS fetch, or curl the page directly. If the price and add-to-cart state are blank, an agent fetcher sees the same blank.

2. 2

   Check your theme's rendering strategy

   Client-side rendered content that populates after page load is invisible to most agent fetchers, which don't execute JavaScript the way a browser does. Server-rendered or statically generated content is the safe default.

3. 3

   Test Core Web Vitals on mobile

   Slow PDPs get deprioritized by both search crawlers and agent fetchers with timeout budgets. See Shopify speed and Core Web Vitals for the agentic era for the specific thresholds that matter now.

4. 4

   Confirm variant data is server-rendered too

   Size and color variants that only populate via a JS-driven dropdown often don't expose their price/availability deltas to a fetcher. Check at least one variant-heavy product.

Section 4 — What should I check before installing a Brand Agent

A Brand Agent is only as good as what you feed it. Shopify's Spring '26 edition shipped Brand Agents on June 17, 2026 — currently Shopify Plus only — and the install is fast, but a fast install on top of thin inputs produces a confident, wrong assistant. Before you turn one on, check these training inputs.

Brand Agent

Shopify's on-site conversational AI assistant, shipped in the Spring '26 edition, that answers shopper questions and can complete transactions using a merchant's catalog, policies, and content as its knowledge source. See Brand Agents explained for the full mechanics.

For the deeper design work — how to actually script tone, escalation, and edge-case handling — see Brand Agent conversation design.

Section 5 — Copilot Checkout edge cases

Copilot Checkout is Shopify's agentic checkout flow, also shipped in the Spring '26 edition alongside Brand Agents, and also currently Shopify Plus only. The core flow — add to cart, confirm, pay — works reliably in a demo. What breaks it in production is the edge cases merchants don't think to test until a real shopper hits them.

• Split shipments. If an order ships from two warehouses or has a backordered line item, does the agent communicate that clearly, or does it imply single-shipment delivery?
• Out-of-stock substitutions. If a variant sells out mid-conversation, does the flow degrade gracefully or dead-end?
• Subscription and recurring products. Agentic checkout flows are built around one-time purchase logic first — confirm subscription products actually complete correctly.
• Address validation failures. Test an address your carrier would flag or correct. Confirm the agent surfaces the correction instead of silently failing the order.
• Discount code interactions. Confirm automatic and code-based discounts apply the way they do in standard checkout — agentic flows have historically lagged the main checkout on promo logic.
• Tax and duty edge cases, if you sell internationally — cross-border orders are where checkout logic is most likely to have gaps.

Section 6 — Measurement

The last checklist item is the one most stores skip: can you actually see agent-driven activity separately from the rest of your traffic? Without it, you can't tell whether Brand Agent conversations convert, whether Copilot Checkout completions are growing, or whether off-site assistants like ChatGPT, Perplexity, and Google AI Overviews are sending you referral traffic at all.

1. 1

   Tag agent and AI-referral traffic distinctly

   Separate on-site Brand Agent sessions, Copilot Checkout completions, and off-site AI-referred sessions into their own reporting buckets rather than lumping them into generic "direct" or "organic" traffic.

2. 2

   Track a conversion funnel specific to agent sessions

   Agent-assisted sessions convert differently than standard browsing sessions. Measuring them with the same funnel hides what's actually happening.

3. 3

   Set a cadence for reviewing AI citation visibility

   Off-site GEO performance — whether assistants cite and recommend your products — changes as your catalog and content change. Check it on a recurring basis, not once.

4. 4

   Define what "working" means before you launch

   Agree on the two or three metrics that will tell you the investment paid off, before you're staring at a dashboard trying to reverse-engineer success after the fact.

For the full framework on what to measure and how to attribute it correctly, see measuring agentic commerce ROI.

What to do with your results

If you ran the six sections above and came back mostly green, you're in a small minority — most stores we assess fail Section 1 outright, which cascades into failures everywhere else, since structured data, Brand Agent accuracy, and GEO citations all read from the same catalog. If you came back with a page of gaps, that's normal, and it's exactly what this checklist is for: finding the gaps before you commit budget, not after.

• If catalog coverage is the gap, start with Catalog Enrichment for AI.
• If structured data and off-site citations are the gap, that's Commerce GEO territory.
• If you're on Shopify and want the whole stack — catalog, structured data, Brand Agent, Copilot Checkout — set up correctly and live fast, that's the Agentic Commerce Setup flagship service.
• If you run an agency and are building this checklist into client delivery, see the agency agentic commerce playbook.