Product catalog readiness

Your product catalog is the foundation of agentic commerce. The quality and completeness of your product data directly impacts product visibility and conversion rates in AI-driven commerce channels.

Product Selection and catalog readiness

Product Selection

Link the Product Selection to the Store because the Store is the assortment boundary used by the channel. This lets you keep one shared catalog in commercetools while exposing only the subset that should be available in that agentic experience.

Common criteria for including or excluding products:

• Brand restrictions: exclude brands or product lines that shouldn't appear in AI search results.
• Inventory thresholds: include only products with sufficient stock levels.
• Data quality: exclude products with incomplete data, missing images, or quality issues.
• Regional availability: limit products to those available in specific markets.
• Promotional focus: create time-limited selections for seasonal or promotional items.

Product Tailoring

Use Product Tailoring when the same Product needs a different presentation in the Store context used by the agentic channel. This is useful when your main storefront content assumes a visual browsing experience, but the agentic channel depends more heavily on descriptive text and direct answer quality.

This is particularly useful because the agentic interface presents products differently than a traditional storefront. For example:

• Titles: include brand name and key differentiators (for example, color) in the title because the agentic interface doesn't have visual UI elements like color swatches.
• Descriptions: incorporate key attributes (for example, waterproof, lightweight) into the narrative text because the agentic interface cannot display custom attributes as structured fields.
• Images: select product shots that work well in text-based and card-based presentation formats.

Product data requirements

AI commerce protocols define fields that describe your products for indexing and discovery. While the exact field names vary by protocol, the data requirements converge on common categories.

Required data

Every product must include the following information for any AI commerce integration:

• SKU identifier for stable product identity across updates.
• Title and description for AI-driven search and natural language matching.
• Product URL and at least one image for visual context and linking back to your site.
• Price and availability status for real-time purchase decisions.
• Inventory quantity for stock-level reasoning.

Strongly recommended data

These fields significantly improve product discoverability and conversion:

• Brand for filtering and differentiation in multi-brand search results.
• GTIN (Global Trade Item Number) or MPN (Manufacturer Part Number) for product identification and validation.
• Product category for proper classification and taxonomy matching.
• Item group identifier for associating product variants (for example, different colors of the same shoe).

Recommended data

These fields enhance the commerce experience and can improve conversion rates:

• Condition (new, refurbished, used) for setting customer expectations.
• Material for filtering and answering composition questions.
• Dimensions (length, width, height, weight) for size-based filtering and shipping calculations.
• Color and size for variant presentation.
• Age group and gender for audience-specific targeting.
• Additional images, video links, and 3D model links for rich media presentation.
• Sale price and sale price effective dates for time-limited promotions.
• Shipping information (delivery estimates, costs) for fulfillment transparency.
• Return policy URL and return window for customer confidence.
• Popularity score and return rate for performance-based ranking signals.

Prioritize fields for your business

Not all optional fields provide equal value for every business. Use this framework to decide which fields to implement:

Implement product attributes

1. Add new Attribute Definitions to your Product Types through the Merchant Center or API.
2. Backfill existing products with correct values before enabling the feed.
3. Mark critical attributes as required in commercetools to enforce data quality.

Product data syndication

Your product catalog needs to be delivered to external AI platforms as a structured data feed. This section explains the concepts involved.

Feed formats and delivery

AI commerce platforms accept product data in structured file formats such as TSV, CSV, XML, or JSON. The data is delivered to a secure endpoint provided by the platform and transferred over HTTPS.

Mapping configurations transform your commercetools product data into the target feed schema. These transformations typically use expression languages such as JSONata to map fields like SKU, title, description, price, and availability from your commercetools data model to the target format.

Initial load and incremental updates

Product data syndication happens in two stages:

1. Initial full load: a complete export of your product catalog establishes the baseline. The AI platform validates the schema and data quality, then indexes all your products. Depending on catalog size, this can take from a few minutes to a couple of hours.
2. Incremental updates (delta feeds): after the initial load, only changed data is synchronized. Changes include price updates, inventory changes, new or removed products, and content modifications. Delta feeds are typically sent at regular intervals (for example, every 15 minutes) to keep the AI platform's index current.

Incremental updates aggregate and deduplicate changes. For example, if a product's inventory changes three times within a single update interval, only the latest state is included in the next delta feed.

Performance considerations

For large catalogs (hundreds of thousands or millions of products), the initial full export is chunked into manageable batches. The commercetools APIs support efficient pagination and batching to ensure the export completes in reasonable time.

Key takeaways

• Product Selections control which products appear in each agentic channel. Link them to your dedicated Store.
• Product Tailoring lets you customize titles, descriptions, and images for the agentic channel without changing your main product data.
• Required product data includes SKU, title, description, URL, image, price, availability, and inventory.
• Strongly recommended fields (brand, GTIN or MPN, category, item group ID) significantly improve discoverability.
• Prioritize optional fields based on customer questions, competitive differentiation, implementation effort, and conversion impact.
• Product data syndication uses structured feeds (TSV, CSV, XML, or JSON) delivered over HTTPS with initial full loads followed by incremental delta updates.

Test your knowledge

On the next page, you will learn how to prepare your checkout flow for agentic commerce.