SendToTeam + Shopify Integration

SendToTeam connects to Shopify through a custom app you install from your Shopify admin. During setup, you grant specific API scopes: read_products and read_orders for reporting and context, write_products if you want AI employees to push approved descriptions back to product listings, and read_customers for personalized post-purchase outreach. Each scope is granted individually, so you control exactly what data flows between the platforms.

Product description writing works by reading your existing catalog data. The AI employee pulls each product's title, category, tags, vendor, price, variant options, and any existing description. It then drafts a new description that incorporates the product's specifications, targets your audience's search queries, and follows a tone you define (conversational, technical, luxury — whatever matches your brand). Approved descriptions can be pushed directly back to the product listing via the Admin API, or exported for manual paste if you prefer.

Order-triggered outreach uses Shopify webhooks. When a customer places their first order (orders/create webhook with order count = 1), the AI employee drafts a welcome email that references the specific product purchased and suggests complementary items from your catalog. When an order ships (fulfillments/create), the draft includes tracking details and product care instructions pulled from your product metafields. For customers who have not ordered in 90 days, a scheduled task generates re-engagement drafts with recommendations based on their purchase history.

Sales reporting aggregates your Shopify order data into summaries your team can act on. A daily report might include total revenue, order count, average order value, top-selling products, and inventory alerts for products below a restock threshold. The assistant queries the Orders API, processes the data, and formats the report for your review queue or a connected Google Sheet.

The integration respects Shopify's API rate limits (2 REST requests per second) by batching and queuing operations. Stores with large catalogs (10,000+ products) may see description generation spread across several hours rather than completing in a single batch.