AI Ecommerce Search: Product Feed Readiness Checklist
Prepare product data for AI search with a field checklist covering identity, variants, attributes, prices, stock, update behaviour and measurement joins.
Product-data requirements for AI ecommerce search
Useful search starts with a catalogue that can answer the shopper’s question and support the purchase that follows. Check product identity, searchable attributes, offer details and update behaviour together. This resource distinguishes API-required fields from the additional data a particular search experience may need.
Start with identity and a valid purchase destination
Every searchable item needs a stable reference. Decide whether it represents a product, a sellable variant or another catalogue entity, then preserve that meaning across feeds and events. A result that points to the wrong size, pack or model creates an avoidable failure even when its title matches the query.
The current PA Products API marks refId, itemGroupId, name, url and isInStock as required. Other fields in that contract are optional, but may be necessary for your intended experience. For example, a price filter needs usable price data even when price is not an API-required field. Current Products API.
Separate searchable meaning from display and eligibility
Descriptions, product type and attributes help retrieval understand the item. Images and titles support presentation. Availability, restricted assortment and verified attribute conditions decide whether an item is eligible. Record which system owns each value and avoid silently substituting generated descriptions for authoritative product facts.
Use typed values. Store weight with an agreed unit, preserve model identifiers as strings and distinguish an absent value from a negative assertion. A text value such as “unknown” is not a useful numeric price. A missing dietary field is not evidence that the product meets a dietary requirement.
Describe variants without losing their relationships
Keep the parent relationship separate from the sellable item identifier. Test groups containing different sizes, colours and pack quantities. Confirm which variant appears on the search card, which stock record it uses and what the destination page selects. Resolve these behaviours before measuring catalogue coverage.
For Shopify, custom product and variant metafields require explicit configuration; changing that configuration also requires a product sync for existing products. Shopify field configuration.
Test updates, not just the first upload
Create a synthetic product, change its name and stock state, and verify what the search experience shows. Then test an item that should no longer be discoverable. Confirm the supported removal or deactivation mechanism with the implementation team rather than inventing a delete operation.
The current Products API warns that omitted fields become empty or null. Treat updates as complete records unless the exact endpoint and deployment contract explicitly permit partial updates. An older ATS overview describes a different update interface, so confirm the version used in your integration. Products API update note.
Join the catalogue to measurement
Verify that result impressions, clicks, basket events and orders refer to the same item identity. Keep a small regression fixture with expected outcomes for additions, changes, missing values and withdrawals. A successful ingestion response is a starting point; visible, purchasable and correctly measured products are the acceptance test.
Illustrative example
Synthetic illustrative Products API record: {"products":[{"refId":"DEMO-OATS-500","itemGroupId":"DEMO-OATS","name":"Example Oats 500g","description":"Synthetic catalogue fixture for ingestion testing.","sku":"DEMO-OATS-500","url":"https://example.com/products/demo-oats-500","isInStock":true,"brand":{"name":"Example Brand","logoUrl":"https://example.com/brand.png"},"images":["https://example.com/oats.png"],"prices":[{"code":"GBP","price":2.5,"discountPrice":2.5}],"contents":[],"attributes":[{"name":"pack_weight_g","scalarValue":500},{"name":"category","stringValue":"Porridge"}]}]}. This example is not live customer data or a live endpoint request.
Downloads
- search-product-feed-checklist.csv (illustrative template)
- search-product-fixture.json (illustrative template)
Sources
- Products API, checked 2026-10-25
- Field Configuration, checked 2026-10-25
- Getting Started With ATS, checked 2026-10-25
How Particular Audience does it
Frequently asked questions
Are all optional API fields optional for our project?
No. API optionality describes the request contract, not every business requirement. A project that needs price filtering, imagery or a particular product attribute must supply usable values for those features and agree how incomplete records are treated.
Can we send only the fields that changed?
Do not assume that. The current Products API says missing fields are cleared. Confirm endpoint version and update semantics with PA, then test a controlled fixture. Preserve existing values when the deployed contract expects complete product records.
Do Shopify metafields sync automatically?
Custom metafields need explicit field configuration, including the correct product or variant scope. The published setup guide also requires a product sync after configuration changes to apply them to existing products. Validate the resulting attributes before relying on them in search.
Related guides
- Store-specific grocery search: stock, assortment and availability: Design grocery search around store assortment, local offers and stock freshness. Explore a worked data model, changed-store behaviour and availability checks.