Agentic Commerce Readiness Checklist: Is Your Product Catalog Ready for AI Shopping Agents?

Agentic commerce is a model of online buying where AI shopping agents search, compare, and purchase products on a customer's behalf, and your store is ready for it only if your product data is structured, your inventory is accurate in real time, and your checkout can complete without a human clicking through it. Most catalogs built for human browsing fail at least three of those tests.

This checklist covers the data, protocol, checkout, and governance work that determines whether agents such as ChatGPT's shopping features, Google's Gemini-based shopping tools, and Perplexity's shopping experience can find, trust, and transact with your products in 2026.

What Agentic Commerce Changes About Your Storefront

A human shopper tolerates vague titles, lifestyle photography, and a hunt for the size chart. An AI shopping agent does not. It converts a request like "waterproof trail shoes under $150, wide fit, delivered by Friday" into structured constraints, then filters catalogs against those constraints. Products with missing attributes are not ranked lower. They are excluded.

That shifts the optimization target from persuasion to machine-verifiable facts. Your copy still matters for human conversion, but agents select from the fields they can parse: price, availability, dimensions, materials, shipping speed, return window, and review data.

Agentic Commerce Readiness Checklist

Work through these in order. Each layer depends on the one before it.

1. Complete product attributes. Every SKU has GTIN or MPN, brand, color, size, material, and condition populated.
2. Schema.org markup. Product, Offer, AggregateRating, and MerchantReturnPolicy are present in JSON-LD on every product page.
3. Merchant feed accuracy. Your Google Merchant Center feed matches on-page price and stock exactly.
4. Real-time inventory. Stock and price updates propagate in minutes, not nightly batches.
5. Machine-readable policies. Shipping times, costs, and returns are structured data, not only footer prose.
6. Agent-compatible checkout. Guest checkout, tokenized payment, and no mandatory CAPTCHA on the purchase path.
7. Bot governance. Your robots.txt and CDN rules distinguish AI agents from scrapers deliberately.
8. Measurement. Agent-referred sessions and orders are tagged and reported separately.

Step 1: Fix Product Data at the Attribute Level

Audit your catalog for attribute completeness, not just product count. A practical method: export all SKUs, then calculate the fill rate for each required field. Aim for 95% or higher on GTIN, brand, color, size, material, and category. Anything below that is a filtering blind spot.

Three problems appear repeatedly in Shopify, BigCommerce, and WooCommerce catalogs:

• Attributes buried in descriptions. "Fits up to 17-inch laptops" in a paragraph is invisible to a filter. Move it to a structured field such as a metafield (Shopify), custom field (BigCommerce), or product attribute (WooCommerce).
• Inconsistent values. "Navy," "Dark Blue," and "Blue-Navy" fragment one color across three values. Standardize to a controlled vocabulary.
• Variant collapse. If every size and color shares one price and one stock figure, agents cannot confirm that a specific variant is purchasable. Each variant needs its own SKU, GTIN, price, and inventory count.

Use Google's product taxonomy for categories. It is the most widely consumed classification, and aligning to it improves matching across multiple agent platforms at once.

Step 2: Implement Structured Data That Agents Can Verify

Agents cross-check. If your JSON-LD says $89.00 and your visible page says $94.00 after a script runs, trust drops and the listing can be suppressed. Render structured data server-side so it matches the initial HTML, and validate it with Google's Rich Results Test after every theme or app change.

Minimum fields for each Offer: price, priceCurrency, availability (using the schema.org InStock or OutOfStock values), itemCondition, and priceValidUntil for sale pricing. Add shippingDetails with deliveryTime and hasMerchantReturnPolicy with return window in days. These two blocks answer the questions agents ask most: how fast and how reversible.

Reviews need the same rigor. Include aggregateRating only when it reflects reviews actually displayed on the page. Mismatches violate Google's guidelines and undermine the signal agents use to rank trustworthy options.

Step 3: Build Feeds and Protocols for Agentic Checkout

Agentic checkout means an agent completes the purchase inside its own interface, passing structured order data to your store rather than driving your checkout page like a browser. Several protocols now support this, including OpenAI and Stripe's Agentic Commerce Protocol, and Google's Agent Payments Protocol (AP2). Platform support varies, so check what your commerce platform exposes before building anything custom.

Plan around three requirements regardless of protocol:

• Tokenized payments. Agents pass a scoped, single-use payment token instead of raw card data. Confirm your processor supports shared payment tokens.
• Deterministic totals. Tax, shipping, and discounts must be calculable through an API call before purchase. If totals only appear after address entry on a rendered page, an agent cannot quote a final price.
• Order status endpoints. Agents report shipping and returns back to the buyer. Expose order status and tracking through API or webhook, not only through emailed links.

The trade-off is control. Agent-completed orders bypass your upsells, cross-sells, and email capture. Offset this by collecting post-purchase consent through the confirmation flow and by pricing bundles as distinct SKUs agents can select directly.

Step 4: Compare Human-Optimized and Agent-Ready Storefronts

Step 5: Set Bot Governance Without Blocking Revenue

Many enterprise stores block automated traffic wholesale through Cloudflare, Akamai, or Shopify's bot protection. That also blocks legitimate agents. Review your robots.txt for user agents such as GPTBot, OAI-SearchBot, PerplexityBot, and Google-Extended, and decide per agent whether to allow it. Note the distinction: some crawlers train models, while others fetch pages to answer a live shopping query. Blocking training crawlers while allowing search and shopping agents is a defensible policy.

At the checkout layer, replace blanket CAPTCHA with risk scoring, and keep fraud rules active. Agent traffic raises new patterns, such as many sessions from shared IP ranges, so tune velocity limits to avoid false declines.

Step 6: Measure Agent-Driven Revenue

Create a dedicated channel grouping in GA4 for referrals from chatgpt.com, perplexity.ai, gemini.google.com, and copilot.microsoft.com. Track add-to-cart rate, conversion rate, and average order value for that group against organic search. Early data typically shows fewer sessions but higher intent, since the agent has pre-filtered the options. Review feed disapprovals in Merchant Center weekly, because a single rejected attribute can remove a product from agent consideration.

Agentic Commerce FAQ

What is agentic commerce?

Agentic commerce is online shopping in which an AI agent researches, compares, and buys products for a customer based on stated preferences. The merchant's structured data and checkout interface determine whether the agent can select and purchase from that store.

What is agentic checkout?

Agentic checkout is a purchase completed by an AI agent through an API-based flow, using a tokenized payment credential and structured order data instead of a human filling out a rendered checkout page.

Do I need to rebuild my store to support AI shopping agents?

No. Most gains come from cleaning product attributes, aligning feeds with on-page data, adding Offer, shipping, and return schema, and enabling the agent-checkout options your platform already supports. Custom API work is only needed for nonstandard catalogs.

Which platforms support agentic commerce today?

Shopify, BigCommerce, and WooCommerce each offer feed, schema, and payment integrations that support agent discovery, with checkout protocol support rolling out at different speeds. Confirm current availability in your platform's documentation before committing to a build.

How do I know if AI agents are sending me traffic?

Filter GA4 referral sources for chatgpt.com, perplexity.ai, gemini.google.com, and copilot.microsoft.com, and review server logs for agent user agents. Compare conversion rate and order value against organic search to see the revenue impact.