Agentic Visibility Checklist v1

Last updated: 2026-05-15

By Jose Galiano · Maelify · maelify.com/book

How to use this

10 checks. Each one gets a Pass / Partial / Fail. Anything not Pass is in your fix queue. The whole audit takes 90 minutes on your top 20 SKUs. You can run it solo.

If the score is below 7/10, your store is probably invisible in agentic surfaces. Book the 20-min Architecture Review at the bottom and we'll triage the priority order.

Section 1 — Catalog truth (4 checks)

☐ 1. Product JSON-LD is present on every PDP

Open DevTools → Elements → search for application/ld+json. Confirm Product schema is rendered server-side (not just injected after page load).

• Pass: Server-rendered Product schema with @type: Product on 100% of PDPs.
• Partial: Present but client-rendered (agents that don't run JS miss it).
• Fail: No Product schema, or only on some products.

☐ 2. JSON-LD includes all 6 must-have fields

For each product page, the schema must include:

• ☐ name
• ☐ image (array, ≥3 URLs)
• ☐ description
• ☐ offers.price + priceCurrency
• ☐ offers.availability (https://schema.org/InStock etc.)
• ☐ gtin13, gtin12, mpn, or sku

• Pass: All 6 on 100% of audited products.
• Partial: 4–5 on most.
• Fail: Fewer than 4 average.

☐ 3. JSON-LD includes review signals

Schema-bearing review data is the single biggest visibility lift in agentic surfaces — agents preferentially recommend rated products.

• ☐ aggregateRating.ratingValue
• ☐ aggregateRating.reviewCount
• ☐ At least 3 individual review entries with author, reviewRating, reviewBody

• Pass: All three on your top 20 SKUs.
• Partial: aggregateRating only, no individual reviews.
• Fail: No review schema even though your site shows reviews to humans.

☐ 4. Inventory and price are real-time-truthful via Storefront API

This is what killed ChatGPT Instant Checkout on March 4, 2026. Agents that pull stale prices via scraping silently drop you.

Open Storefront API console → query a product variant → confirm quantityAvailable and price.amount match what a customer sees on the PDP and at checkout. Within the same second.

• Pass: Storefront API returns identical inventory + price to what the customer sees, every time.
• Partial: Matches most of the time but drifts during high-traffic moments.
• Fail: Storefront API returns null inventory, or a different price than checkout.

Section 2 — Discovery experience (3 checks)

☐ 5. Test discovery in all 4 agents — same query format

Pick your top 5 SKUs. Run this query format in ChatGPT, Perplexity, Gemini, Copilot:

For each agent, record: Did your product appear in top 3? Were the details (price, variant, availability) correct?

• Pass: Top 3 in 3+ agents, details correct.
• Partial: Surfaces but with wrong details or only in 1–2 agents.
• Fail: Doesn't surface, or shows a competitor where you should be.

☐ 6. Variant disambiguation

Run this query in ChatGPT: "buy [your specific variant name and option, e.g. blue size large] from [your brand]".

Did the agent identify the correct variant? Or did it return a different color/size/flavor?

• Pass: Exact variant returned every time.
• Partial: Right product, wrong variant.
• Fail: Wrong product entirely.

☐ 7. Geography and shipping rules

Run: "recommend [your product] for someone in [country you ship to]".

If you ship there, you should appear. If you don't ship there, you should not appear (false positives erode agent trust over time).

• Pass: Geography accuracy matches your actual ship-to footprint.
• Partial: Surfaces in some markets but not others where you do ship.
• Fail: Surfaces where you don't ship, or hides where you do.

Section 3 — Catalog structure (3 checks)

☐ 8. Metafields cover structured attributes

For each product, you should have metafields covering:

• ☐ Use case (e.g. "morning routine", "post-workout", "travel")
• ☐ Audience (e.g. "men 25-45", "B2B buyers", "gift buyers")
• ☐ Material / ingredient list (structured, not paragraph text)
• ☐ Dimensions / weight / pack size
• ☐ Compatibility (for tech/accessory products)

• Pass: 4 of 5 categories filled with structured data on top 20 SKUs.
• Partial: Some categories but mostly as free text in the description.
• Fail: No metafields beyond what Shopify auto-generates.

☐ 9. Product taxonomy mapping is current

Shopify's standard product taxonomy was overhauled in 2025. If your products are mapped to deprecated category IDs, agents struggle to match them to user intent.

Check: Admin → Products → bulk view → "Product type" and "Category" fields.

• Pass: All audited products mapped to Shopify's current standard taxonomy.
• Partial: Some products on legacy categories.
• Fail: Custom-string product types, no standard category mapping.

☐ 10. Catalog API exposure verified

Shopify ships the Catalog API on every Plus store, but enabling syndication to AI partners requires a confirmation step in Admin. Many Plus stores haven't done it.

Check: Admin → Sales channels → confirm "AI shopping channels" or equivalent is enabled (UI name may have shifted post-Winter '26 Edition).

• Pass: Confirmed enabled, last sync within the past 24 hours.
• Partial: Enabled but last sync >7 days old.
• Fail: Not enabled, or no record of last sync.

Scoring

• 9–10 Pass: You're in the top 10% of Shopify stores for agentic visibility. Keep auditing weekly.
• 6–8 Pass: You'll surface intermittently. Fix your Fails first — usually JSON-LD and inventory-truth are the bottleneck.
• 0–5 Pass: You're effectively invisible. This is the typical state of most Shopify stores in May 2026. Don't feel bad. Fix it.

The 3 fixes that move the most

If you only have one engineering sprint and you have to triage:

1. Get all 6 required JSON-LD fields on every PDP, plus aggregateRating with real review data.

2. Wire your inventory and price to a single source of truth, exposed via Storefront API.

3. Add 4–5 metafields per product covering use case, audience, material, and dimensions.

That's roughly 1.5 weeks of work for one senior engineer on a typical Plus catalog. It's not architecture. It's hygiene. And it's worth ~3x your agentic discovery rate.

When to call me in

If your score is below 6/10 and your team doesn't have the engineering bandwidth to fix it in the next 30 days, book a 20-min Architecture Review and we'll scope it together. I run a 2-week Scoping Sprint that delivers:

• A full audit of your top 100 SKUs against this checklist
• A prioritized punch list with effort estimates
• A reference implementation of the top 3 fixes on your top 20 SKUs
• A weekly discovery test rig your team can run after I'm gone

Changelog

• v1 (2026-05-15): Initial release. 10 checks across catalog truth, discovery, and structure.

This checklist is updated as the agentic commerce protocols evolve. Current as of May 2026 — covers UCP March '26 multi-item update, ACP via ChatGPT discovery (post-March-4 Instant Checkout removal), Agentic Storefronts default rollout (March 24, 2026), and Shopify Catalog API public release.