Identify the exact item
Keep product IDs stable. Supply the actual brand and assigned identifiers when applicable; do not invent data to improve a score.
PRODUCT DATA FOR AI SHOPPING
Is your product catalog ready for AI shopping?
See where your structured product data is complete and where useful context is missing. The live scanner returns a separate, explainable AI Shopping Readiness score beside your General Feed Score.
Check AI shopping readinessIndependent framework v1. No guarantee of ranking, recommendations, visibility or sales in ChatGPT, Gemini or any other system.
What is AI shopping readiness?
AI Shopping Readiness is an independent catalog-data assessment, not certification from an AI provider. It measures observable coverage and formatting proxies that can make product records easier to interpret: identity, descriptive content, category context, declared variants, assets and optional enrichment.
A complete field can still be inaccurate. This tool does not establish factual truth, test a commerce agent or predict whether a product will appear in a conversation.
What product data helps describe your catalog?
Explain what it is
Use factual titles, descriptions and category context. Length and presence checks are useful signals, not a substitute for editorial review.
Describe meaningful differences
Where relevant, represent variant groups and actual options, plus useful images and structured product details.
Six dimensions, one separate readiness assessment
- Identity
- Content
- Taxonomy
- Variants
- Assets
- Enrichment
Open each dimension in your scan result to see its score, evaluated products, gaps and recommended priority. Baseline gaps, strong recommendations and advanced enrichment are labelled separately. Missing advanced data is not automatically a critical error.
Variants are only assessed when grouping is declared. An unobserved group is not proof that the store has no variants.
A clean feed can still have enrichment gaps
| Measurement | Example result | Interpretation |
|---|---|---|
| General Feed Score | 100 / 100 | No failures in the general rules |
| AI Shopping Readiness | 88.89 / 100 | Optional enrichment is absent |
| Variants | Not applicable | No grouping metadata was declared |
This example has valid general data but no optional color/material, additional image or product detail. The two frameworks remain separate; the readiness gap does not convert a passing general check into a failure.
Common catalog gaps to review
Missing identity or category
Repair source mappings and confirm identifiers and categories against real product records. Recognized identifiers are not applicable to every product.
Thin or damaged descriptions
Review factual copy upstream. Recovered encoding characters are not treated as useful text, and the checker does not fabricate replacements.
Unclear options or optional context
Inspect declared variant groups and useful attributes. Do not add irrelevant enrichment merely to raise the score.
Feed compliance is not AI shopping readiness
A destination-specific validator checks a channel schema. This broader assessment asks which supported pieces of catalog context are present. It cannot certify any provider's feed or infer organic recommendation decisions.
Preparing native OpenAI ads data? Use the separate ChatGPT Ads Feed Checker. For general source-quality priorities, read Product Feed Quality and Product Feed Audit.
What the score measures - and what it does not
Version 1 applies deterministic checks to the same normalized stream used by the general audit. Each dimension measures affected versus evaluated products. The overall result is a weighted mean of applicable dimensions; non-applicable dimensions are excluded. General audit weights are not changed.
We do not crawl product or image URLs, check image dimensions, judge semantic accuracy, reconcile variants across rows, inspect external documentation or query AI providers. The result is diagnostic guidance, not proof of platform acceptance.
Some input names follow Google product-data conventions. They are not a universal standard for AI agents. The private PDF repeats the framework and its limitations.
Framework review date: .
AI product feed readiness questions
VISIBLE FAQ
Questions about this guide
What is AI shopping readiness?+
It is Product Feed Scan's independent, versioned assessment of data-coverage proxies that can help a catalog describe its products more clearly. It is not a universal AI standard or a test of how a specific model understands a product.
Will a high score get products recommended by ChatGPT or Gemini?+
No visibility, ranking, recommendations or sales are guaranteed. The checker does not query those systems, measure their recommendations or predict their decisions.
Is this the same as channel feed compliance?+
No. A channel validator checks a specific destination's schema. AI Shopping Readiness uses broader catalog-quality proxies. Use the separate native ChatGPT Ads checker when preparing that specific feed format.
Why are General Feed Score and AI Shopping Readiness different?+
They apply different documented frameworks to the same parsed source. General rules remain unchanged. AI readiness groups coverage into six dimensions and treats missing optional enrichment as a suggestion, not a critical error.
Are missing variants always a problem?+
No. Without declared grouping metadata, the variant dimension is marked not applicable. That does not prove variants are absent; review the source mapping if your catalog contains them.
Which input formats and sizes are supported?+
Use a public XML or CSV URL or temporary XML/CSV upload. Transfer and decompressed data are each limited to 200 MiB, with at most 100,000 products. No store login or authenticated feed URL is needed.
What do I get for free and in the full report?+
The free scan includes both scores, dimension summaries and the top three general issue groups. The €19 one-time report adds complete general findings, product-level repair exports, PDF and email delivery for 30 days. AI readiness remains an aggregate dimension assessment; it does not add a second set of product-level CSV findings.
Inspect your catalog before adding more data.
Public XML/CSV or temporary upload. Free scores; complete general report €19 once, with 30-day access.