Skip to content
UCP
Menu

Playbook

Merchant readiness checklist

A 47-point audit to determine whether your catalog, feeds, policies and transactional stack are ready for AI agents. Work through it with your PIM, feed and merchandising owners.

Updated : April 2026 · Primary query : merchant readiness ai agents

This is an operator-grade audit. Each item has a pass criterion and a leverage note. Items marked P0 are blocking — if they fail, later items are moot. Items marked P1 meaningfully improve agent performance. P2 items are polish. Target: >80% of P0 and P1 passing before you consider yourself agent-ready.

Identity (P0)

  1. GTIN present on >95% of SKUs where applicable.
  2. MPN + Brand present on SKUs without GTIN.
  3. Brand field normalized (single canonical spelling per brand).
  4. SKU stability — internal SKU never reused for a different product.
  5. Variant identity — size/color variants have distinct GTIN or a stable variant ID.

Addressability (P0)

  1. Canonical URL tag on every PDP.
  2. First-byte HTML includes product name, price, brand — not waiting for client-side hydration.
  3. No hard consent wall blocking first-request rendering for crawlers / agents.
  4. XML sitemap includes all active PDPs.
  5. robots.txt explicitly allows major AI crawlers you want indexed.

Semantic coverage (P1)

  1. Structured data: JSON-LD Product + Offer on every PDP.
  2. Category placement: Google Product Category + GS1 GPC where relevant.
  3. Typed attributes covering category query surface (material, dimensions, technical specs).
  4. Units explicit (metric + imperial where relevant).
  5. Enumerated values where applicable (sizes, colors from a known list).
  6. Localized values: size conversions, currencies, units.
  7. Image set: primary, gallery, role-tagged (front, back, detail, lifestyle).
  8. Alt text: descriptive, not decorative.

Offers and freshness (P0–P1)

  1. Price field in feed and JSON-LD, with currency.
  2. Promotion price fields correct, with validity window.
  3. Availability state (in_stock, out_of_stock, pre_order, back_order).
  4. Stock quantity exposed where meaningful.
  5. Price parity: feed vs PDP vs cart >99% on sampled SKUs.
  6. Stock parity: feed vs checkout >98%.
  7. Feed refresh cadence documented and met.
  8. updated_at timestamps present and accurate.

Policies as data (P1)

  1. Returns policy expressed as MerchantReturnPolicy JSON-LD.
  2. Shipping policy expressed as ShippingRateSettings or equivalent.
  3. Warranty terms structured (duration, scope, claim method).
  4. Age / geographic restrictions machine-expressible where applicable.
  5. Subscription terms if any: frequency, cancellation policy, substitution rules.
  6. Returns landing page still exists for humans in addition.

Discoverability for agents (P1)

  1. Merchant Center feed clean: no high-severity errors.
  2. Marketplace listings consistent with own-site catalog (identifier-wise).
  3. Indexing parity across Google, Bing, and emerging AI crawlers.
  4. Structured data testable via Rich Results / Schema Markup Validator.
  5. Breadcrumbs structured (BreadcrumbList JSON-LD).
  6. FAQ blocks on high-intent pages with FAQPage JSON-LD.

Transactional readiness (P1–P2)

  1. PSP supports agent-pay (Stripe ACP, Visa IC, Mastercard Agent Pay, PayPal agent — whichever is relevant).
  2. Order lifecycle events emitted as webhooks.
  3. Agent user-agent detection in logs and analytics.
  4. Agent-initiated returns supported via API or dedicated endpoint.

Trust and identity (P2)

  1. Merchant identity page (About, legal, contact) with structured data.
  2. Reviews structured (AggregateRating + Review JSON-LD) and honest.
  3. HTTPS universal, HSTS set.
  4. Certificates / labels exposed as structured claims where relevant (organic, fair-trade, etc.).

Observability (P1)

  1. Agent traffic dashboard: crawl frequency per major UA, PDPs fetched, feed pull frequency, conversions attributed to agent-driven sessions.

Scoring

Pass rate on P0Pass rate on P0+P1Readiness
<80%—Not ready. Address P0 first.
≥80%<70%Partial. Visible to some agents, degraded for many.
100%≥85%Agent-ready. Monitor and iterate.
100%≥95% + P2 >50%Agent-optimized. Competitive advantage.

How to work through it

  1. Sample 50 SKUs across your top 5 categories.
  2. Score each SKU against the P0 + P1 items.
  3. Aggregate by category; identify categories with <80% pass rate.
  4. Build a 90-day remediation plan with PIM, feed and engineering owners.
  5. Re-run quarterly. Track pass-rate as a board-level metric.

For a deeper methodology, see audit methodology.

Where to go next