Scout: Agentic Commerce Catalog Auditor

AI agents audit your product feed against the OpenAI Agentic Commerce Protocol and Google Universal Commerce Protocol specs, score its legibility to shopping agents, and draft compliant rewrites

Problem statement

As AI shopping agents replace keyword search, products with incomplete or unstructured catalog data become invisible: an agent cannot list, match, or transact what it cannot read. Scout audits a product feed against the live agentic-commerce specs (OpenAI Agentic Commerce Protocol, Google Universal Commerce Protocol), scores its legibility 0 to 100, flags the highest-impact missing or malformed fields, and drafts compliant rewrites for the worst SKUs, then re-validates its own output. The demo runs on a synthetic catalog.

Run Demo audits the sample catalog against the OpenAI Agentic Commerce Protocol spec.

Run Demo (Alternate Catalog) audits a second synthetic catalog against the Google Universal Commerce Protocol spec, so you can see the same catalog scored against a different spec.

Sample catalog & spec

Input: a product-feed CSV (one row per SKU), or a public Shopify storefront /products.json URL

Per SKU: id, title, description, link, image, availability, price, brand, GTIN, plus recommended fields (category, condition, color, size, …)

Spec: the OpenAI Agentic Commerce Protocol and Google Universal Commerce Protocol required + recommended field schema (bundled snapshot, re-validated against the live version at run time)

Score: a transparent 0 to 100 legibility rubric weighted by each field's impact on whether an agent can list and sell the item