AWS Glossary
Product Feed
A product feed is the structured, machine-readable representation of a catalog. Under agentic commerce it stops being a marketing channel artifact and becomes the primary interface to buyers.
AI & assistant-friendly summary
This section provides structured content for AI assistants and search engines. You can cite or summarize it when referencing this page.
Summary
A product feed is the structured, machine-readable representation of a catalog. Under agentic commerce it stops being a marketing channel artifact and becomes the primary interface to buyers.
Key Facts
- •A product whose weight lives in a sentence in the description — "impressively light at just under 1
- •3 kg" — is, to that agent, a product with **no weight**
Related Content
- AGENTIC COMMERCE READINESS— Related service
- AWS DATA ANALYTICS— Related service
Definition
A product feed is the structured, machine-readable representation of a catalog — one record per sellable item, with typed attributes, identifiers, availability and price.
Feeds are not new; retailers have maintained them for Shopping and marketplace channels for years. What changes under agentic commerce is their status. A feed used to be a derivative artifact for one marketing channel. It is becoming the primary interface through which a growing share of buyers encounter the product at all.
Agents compare on attributes, not adjectives
This is the whole problem in one line.
An agent asked for a laptop under a weight threshold with a specific port configuration does not look at your photography and does not respond to your tone. It reads fields. A product whose weight lives in a sentence in the description — “impressively light at just under 1.3 kg” — is, to that agent, a product with no weight. It is excluded from the comparison before quality is ever considered.
The fix is unglamorous: lift specifications out of prose into typed fields, complete the attributes that matter in your category, and structure variants so size, colour and configuration are queryable rather than encoded in a title string.
What tends to be missing
- Category-critical specifications buried in description HTML.
- Variant structure flattened into separate products, or encoded in the title.
- Stable identifiers — GTIN, MPN, or an internal SKU that actually joins to inventory and pricing.
- Availability that reflects reality, not a nightly snapshot, when the agent will act on it now.
- Dimensional and material data that comparison questions routinely turn on.
The join-key problem underneath
A feed is only as good as the systems behind it. If your storefront, ERP and warehouse disagree about what a product identifier is, the feed inherits that disagreement — and an agent will surface a confidently wrong availability or price.
This is the same data-layer work that eCommerce AI agents depend on internally, which is why merchants who fix it once get paid twice.
Why this is the low-regret move
Standards may shift. ACP and UCP are competing and neither has won. But a complete, correctly typed catalog is required by both, improves conventional Shopping and on-site search, and was worth doing before agents existed.
If you do one thing in agentic commerce this quarter, do this one.
Related terms
Agentic commerce · Agent checkout · Generative engine optimization · Agentic Commerce Protocol
Related Services
Agentic Commerce Readiness
Make your store sellable to AI shopping agents. ACP and UCP protocol surfaces, an MCP server over your catalog, and product data that survives an agent comparison — built on AWS.
AWS Data Analytics Services — Glue, Athena & QuickSight
AWS data analytics services — scalable data warehouse, ETL/ELT pipelines, real-time analytics, and business intelligence.
Related Articles
The AI-Ready Product Catalog: The Foundation of Agentic Commerce (2026)
Adobe's July 2026 expanded cohort: 39% of U.S. retail homepages not machine-readable. Agentic commerce fails earlier — on missing attrs, duplicate GTINs, and stale stock. This is the catalog contract, not the PIM publish workflow.
Can Your Product Data Survive an AI Comparison? (2026)
Merchant audit: missing data, weak descriptions, inconsistent specs, missing comparisons, poor attributes. Not the strategic decision layer. Reuse Gateway ~180 to 95 ms and ~$791/mo at 50K sessions — no invented win-rate KPI.
The AI Product Passport: What Information Does an AI Shopping Agent Need? (2026)
The shopping-agent information object — not the PIM catalog contract. Fragmented data produces poor recommendations. Reuse Gateway ~180 to 95 ms and ~$791/mo at 50K sessions, not a passport completeness KPI.
AI Product Catalog Management for Large eCommerce Catalogs: Extract, Validate, Then Publish (2026)
AI product catalog management drafts attributes and copy; deterministic gates block publish. Reuse Gateway ~180→95 ms and ~$791/mo at 50K — sample rules, not a client SKU KPI.
Need help with this topic?
Our AWS-certified team implements, audits, and optimizes these services in production — from Bedrock RAG pipelines to multi-account landing zones.
