← Back to Resources
Guide

Make product pages legible to AI shopping agents

Lantern Team·11 min read

An AI-legible product page is one a model can read, summarise, and cite without reaching for somebody else's words about you. Most DTC PDPs in Lantern's 200-brand sample do not meet that bar. Across the sample, only 47% of tracked PDPs in the bottom quartile carry Product schema; 91% of top-quartile PDPs do.

Product schema — the floor

Every product page needs Product schema, every variant needs to be exposed, and every claim that has a proof needs to be machine-readable. The minimum viable schema for a coffee subscription PDP:

json{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Ethiopia Yirgacheffe — Whole Bean, 12 oz",
  "brand": { "@type": "Brand", "name": "Wildgrain Roasters" },
  "sku": "WG-ETH-YIR-12",
  "description": "Direct-trade single-origin from the Yirgacheffe region. Bright citrus and jasmine. Roasted-to-order on Tuesdays.",
  "offers": {
    "@type": "Offer",
    "price": "21.00",
    "priceCurrency": "USD",
    "availability": "https://schema.org/InStock"
  },
  "additionalProperty": [
    { "@type": "PropertyValue", "name": "originCountry", "value": "Ethiopia" },
    { "@type": "PropertyValue", "name": "tradeBasis",    "value": "Direct-trade" },
    { "@type": "PropertyValue", "name": "brewMethodsBest", "value": "pour-over, AeroPress" }
  ],
  "aggregateRating": {
    "@type": "AggregateRating",
    "ratingValue": "4.8", "reviewCount": "412"
  }
}

FAQ schema

json{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "Is this coffee direct-trade?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes. Wildgrain buys this lot direct from the Konga washing station in Yirgacheffe; the trade premium is published on the origin page."
      }
    },
    {
      "@type": "Question",
      "name": "What brewing methods work best for Ethiopia Yirgacheffe?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Pour-over (V60 or Kalita Wave) and AeroPress bring out the citrus and jasmine notes."
      }
    }
  ]
}

Google Shopping feed attributes

High-leverage Google Shopping feed attributes for DTC.
Feed attributeWhy it mattersCommon gap
google_product_categoryRoutes the SKU to the right buyer-intent clusterMapped to top-level category, not leaf
materialFilters answers on physical attributesBlank on apparel/home goods
gift_card / gift_appropriateCaptures gift-intent promptsMost brands skip
sustainability_certificationFilters answers on values-based promptsBrands have certs, but not in feed
additional_image_linkDrives carousel selectionOne image only

Hero copy — the first 200 words

BrightEdge's finding that 44.2% of all LLM citations come from the first 30% of a page's text is the most actionable single sentence in the AI-search literature. On a typical 800-word PDP, that is the first ~240 words. Whatever you most need the model to lift goes there — not at the bottom, not behind a tab.

The PDP checklist
Product schema with all category-relevant additionalProperty attributes · FAQ schema on 3+ buyer-intent questions · aggregateRating with reviewCount on every SKU with > 10 reviews · Google Shopping feed at 18+ of 28 attributes · hero copy that echoes buyer-prompt language in the first 200 words · every claim paired with a proof artefact · comparison hook in the hero where competitor citations appear.