Home › Blog › Shopify product FAQ page schema

Shopify product FAQ page schema for AI shopping agents: the pre-purchase query playbook

2026-06-10 — ~15 min read — FAQPage schemaAI shoppingJSON-LDShopify Liquid

AI shopping agents answer pre-purchase questions — "is this BPA-free?", "does it fit a king mattress?", "what’s the warranty?" — from structured data on your product pages. Without FAQPage JSON-LD, agents either skip the question or extract an answer unreliably from description text. The difference is whether your store gets cited or your competitor’s does.

Why pre-purchase queries are the silent conversion lever

The AI shopping funnel is different from the Google Shopping funnel. In Google Shopping, a customer sees your product image, price, and title in a grid. They click through to verify the details themselves. The product page carries the conversion burden.

In AI shopping — ChatGPT Shopping, Perplexity Shopping, Google AI Mode — the funnel is reversed. The agent answers the question first, then surfaces a product recommendation. The customer’s pre-purchase questions are resolved before they ever see a product page. If the agent can’t answer the question confidently from your structured data, it either skips your product or cites a competitor whose catalog answers the question clearly.

This is the gap that FAQPage JSON-LD fills. A FAQPage block on a product page is a machine-readable answer bank. Each Question/Answer pair is a direct, attributable answer the AI agent can cite with confidence. There’s no parsing ambiguity — the question text and answer text are explicit fields.

How AI agents use FAQPage JSON-LD

FAQPage is a schema.org type that wraps a list of Question and Answer pairs. Each Question has a name (the question text) and an acceptedAnswer property containing an Answer with a text field (the answer text).

When an AI shopping agent crawls a product page, it indexes FAQPage entries separately from the Product entity. The question text becomes a match target for conversational queries. If a user asks "is this water bottle BPA-free?" and the crawled page has a Question with that text and a clear Answer, the agent has a high-confidence source to cite.

The key distinction from product descriptions is attribution confidence. When an agent extracts an answer from paragraph text, it is making an inference — the answer might be there, might be approximate, might conflict with text elsewhere on the page. When the agent reads a Question/Answer pair, the merchant is making an explicit claim. The agent can cite it as a direct merchant statement, not an inferred fact.

Google uses FAQPage to generate rich results in standard search (expanded Q&A beneath the main result). But in AI Mode and Perplexity, the use is different: FAQ pairs are used to populate the conversational answer layer — the part of the response that directly answers the user’s question before links are shown.

The 5 question categories that cover 80% of pre-purchase queries

Pre-purchase queries follow consistent patterns across product categories. Based on AI shopping query patterns, these five categories account for the majority of pre-purchase questions that agents are asked before a recommendation is made:

For each product, pick 3–6 questions from the categories most relevant to that product type. A kitchen appliance might need compatibility + materials + warranty. A clothing item needs sizing + returns. More than 8 questions per product creates diminishing returns and risks diluting the most important answers.

Implementation: the two-block architecture on Shopify

Shopify themes output Product JSON-LD in a single <script type="application/ld+json"> block. FAQPage lives in a second, separate block on the same page. Do not merge them unless you are using @graph (covered in the @graph section below).

The basic FAQPage block for a product page looks like this:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "Is this water bottle BPA-free?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes. The bottle body and lid are made from BPA-free Tritan co-polyester. The silicone gasket is food-grade and phthalate-free. All materials are FDA-compliant."
      }
    },
    {
      "@type": "Question",
      "name": "Is this dishwasher safe?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "The bottle body is top-rack dishwasher safe. The lid should be hand-washed — high dishwasher temperatures can degrade the silicone seal over time."
      }
    },
    {
      "@type": "Question",
      "name": "What is the warranty?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "This bottle carries a lifetime guarantee against manufacturing defects. If the lid or seal fails under normal use, we replace it free of charge. Contact support@example.com with your order number."
      }
    }
  ]
}
</script>

Three structural rules to follow:

1. Question name must be a question. It should end with a question mark and be phrased as a customer would ask it. Not "BPA status" — "Is this BPA-free?"
2. Answer text must be a complete sentence. Fragment answers ("Yes") are valid JSON-LD but are too thin to be cited confidently by AI agents. Full, specific answers win more citations.
3. No HTML in answer text. The text field is a plain string. HTML entities like & will be parsed as literal characters. Keep answers in plain prose.

Metafield architecture for scalable FAQ at catalog level

Hard-coding FAQ questions into a product template only works for one product. For a real catalog, you need the question/answer content stored in Shopify metafields so each product has its own FAQ data, editable from the admin.

The recommended architecture uses a single JSON metafield per product:

Store the metafield value as a JSON array:

[
  {
    "question": "Is this BPA-free?",
    "answer": "Yes. The body is Tritan co-polyester, the lid and seal are food-grade silicone. All materials are FDA-compliant and phthalate-free."
  },
  {
    "question": "Is this dishwasher safe?",
    "answer": "The bottle body is top-rack dishwasher safe. Hand-wash the lid to preserve the silicone seal."
  },
  {
    "question": "What is the capacity?",
    "answer": "This bottle holds 32 oz (946 ml). The wide mouth is 2.2 inches in diameter and fits standard ice cubes."
  }
]

Then in your product.liquid or sections/main-product.liquid:

{% assign faq_json = product.metafields.product_faq.questions.value %}
{% if faq_json and faq_json.size > 0 %}
<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {% for item in faq_json %}
    {
      "@type": "Question",
      "name": {{ item.question | json }},
      "acceptedAnswer": {
        "@type": "Answer",
        "text": {{ item.answer | json }}
      }
    }{% unless forloop.last %},{% endunless %}
    {% endfor %}
  ]
}
</script>
{% endif %}

Key points about this Liquid pattern:

• The {% if faq_json and faq_json.size > 0 %} guard ensures the FAQPage block only renders for products with FAQ data — no empty mainEntity: [] blocks.
• The | json filter escapes quotes, backslashes, and special characters in question and answer strings, preventing JSON parse errors from apostrophes in answers.
• product.metafields.product_faq.questions.value returns the parsed array (not the raw JSON string) when the metafield type is set to JSON in the admin.

Before/after: a water bottle product

Before: description-only product page

A typical Shopify product page for a water bottle has a Product JSON-LD block generated by Dawn or the active theme. It contains name, description, offers, possibly an aggregateRating. The description field contains prose like:

"Our 32oz BPA-free Tritan water bottle is perfect for the gym, hiking, and everyday hydration. Features a leak-proof lid and wide mouth for easy filling. Dishwasher safe. Backed by our lifetime guarantee."

An AI agent asked "is this water bottle BPA-free?" can attempt to extract this from the description. But the extraction is uncertain: Is "BPA-free" in the title, the description, a metafield, a product tag? The agent makes an inference and may hedge the answer (“the listing says it’s BPA-free”), or flag uncertainty, or skip the product entirely if it finds conflicting information on the page.

After: with FAQPage JSON-LD

The same product page now has a second JSON-LD block:

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "Is this water bottle BPA-free?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes. The bottle body is made from BPA-free Tritan co-polyester. The lid and silicone seal are food-grade and phthalate-free. All materials are FDA-compliant."
      }
    },
    {
      "@type": "Question",
      "name": "What is the capacity of this water bottle?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "32 oz (946 ml). The wide mouth is 2.2 inches in diameter and fits standard ice cubes."
      }
    },
    {
      "@type": "Question",
      "name": "Is this water bottle dishwasher safe?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "The bottle body is top-rack dishwasher safe. Hand-wash the lid to preserve the silicone seal over time."
      }
    },
    {
      "@type": "Question",
      "name": "What is the warranty on this bottle?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Lifetime guarantee against manufacturing defects. If the lid or silicone seal fails under normal use, we replace it free of charge."
      }
    }
  ]
}

Now when an AI agent is asked "is this water bottle BPA-free?" it has a direct, attributable answer. No inference required. The agent can say: "Yes, according to the product page, it is BPA-free Tritan with a food-grade silicone seal." Confident citation. The store gets surfaced; the competitor with only description-text coverage does not.

Integrating FAQ into a full @graph with Product

If you are building a custom product JSON-LD block from scratch (rather than relying on the theme’s auto-generated block), you can consolidate Product and FAQPage into a single @graph array. This is cleaner than two separate script tags and allows you to establish an @id relationship between the FAQPage and the Product page URL.

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "Product",
      "@id": "https://yourstore.myshopify.com/products/32oz-tritan-water-bottle#product",
      "name": "32oz Tritan Water Bottle",
      "description": "BPA-free 32oz Tritan water bottle with leak-proof lid. Top-rack dishwasher safe body. Lifetime guarantee.",
      "brand": { "@type": "Brand", "name": "YourBrand" },
      "offers": {
        "@type": "Offer",
        "priceCurrency": "USD",
        "price": "34.99",
        "availability": "https://schema.org/InStock",
        "url": "https://yourstore.myshopify.com/products/32oz-tritan-water-bottle"
      }
    },
    {
      "@type": "FAQPage",
      "@id": "https://yourstore.myshopify.com/products/32oz-tritan-water-bottle#faq",
      "mainEntity": [
        {
          "@type": "Question",
          "name": "Is this water bottle BPA-free?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Yes. Tritan co-polyester body, food-grade silicone seal. All materials are FDA-compliant and phthalate-free."
          }
        }
      ]
    }
  ]
}
</script>

The @id on the FAQPage entity uses the product URL with a #faq fragment, disambiguating it from the Product entity. This is optional but recommended when using @graph — it makes the entity graph unambiguous for validators.

In Liquid, generate the product @id using {{ shop.url }}{{ product.url }} and append the fragment. The rest of the graph follows the same metafield-driven pattern described in the implementation section above.

5 FAQPage mistakes on Shopify product pages

FAQ