Product data for AI agents: GTINs, feeds and the fields that matter
Agents do not read your product descriptions. They read identifiers, attributes, availability and price. Here is the product data checklist every retailer needs.
Your copywriter spent three weeks on the Trail Runner Pro description. The agent read your JSON in eleven milliseconds and moved on because availability was null. Product data is not a back-office concern for agentic commerce — it is the storefront.
The agent-readable product record
Think of your product record as an API response, not a page. Every field should be structured, validated, and consistent across every surface where agents might read it.
| Field | Priority | Agent use |
|---|---|---|
| gtin / ean | Critical | Identity and cross-retailer matching |
| availability | Critical | In-stock filter and delivery feasibility |
| price + currency | Critical | Budget filter and comparison |
| brand | High | Preference matching and authenticity |
| shipping / delivery | High | Date-specific intent (e.g. by Friday) |
| aggregateRating | High | Trust scoring |
| returnPolicy | Medium | Risk reduction for agent recommendation |
| color / size / material | Category-dependent | Spec matching |
Feed architecture that agents trust
- PIM as single source of truth — no manual spreadsheet exports.
- Automated feed generation to Google, Amazon, marketplaces, and open catalog endpoints.
- Validation rules that reject records with missing GTIN or null availability.
- Monitoring that alerts on feed age, error rate, and PDP mismatches.
- Category-specific attribute templates (footwear, electronics, grocery, etc.).
On-site structured data
Feeds alone are not enough. Agents and crawlers also read your PDP. Implement Schema.org Product with Offer, AggregateRating, and Review where applicable. Use JSON-LD in the page head. Validate with Google's Rich Results Test and fix warnings — not just errors.
Frequently asked questions
- What product data do AI agents need?
- At minimum: GTIN, brand, title, description, price, currency, availability, condition, images, category, and delivery/return attributes. Agents also use structured reviews, MPN, colour, size, and material where relevant to the category.
- Why is GTIN so important for agentic commerce?
- GTIN is the global product identifier agents use to deduplicate and cross-reference listings across retailers, marketplaces, and comparison engines. Without it, your product may not match customer intent or may be treated as a duplicate listing.
- How often should product feeds update?
- For fashion and high-velocity categories, aim for near-real-time stock and price updates (minutes, not days). Agents penalise stale feeds — a price or availability mismatch between feed and live site is a trust failure.
Find out what agents see on your estate.
Fixed-scope Agentic Commerce Readiness Audit. £1,950. Ten working days. Scored report, prioritised roadmap, 60-minute debrief.
Book the 30-minute call ↗