Automate Product Description Drafts From a Catalog CSV
Set up a catalog workflow that takes one product row and returns a draft keyed by SKU. Use the product data as the source for every size, material and compatibility claim, then run a test batch before anything reaches your storefront.
Worked example
Example catalog row and checked draft
SKU = BAG-014-OLIVE Item = Packing cube Color = Olive Material = Ripstop nylon Dimensions = 26 × 18 × 8 cm Care = Not provided Draft = An olive packing cube in ripstop nylon, measuring 26 × 18 × 8 cm. Source fields = item, color, material, dimensions Review result = Care claim omitted. SKU and units match the source.
Steps
- Export a few catalog rows with unique SKUs, variant names, materials, dimensions and approved care details. Preserve the units alongside the values.
- Map each row into the generator and ask for its exact SKU, a short draft and the list of source fields used. Omit missing facts and route products with an empty required field to review.
- Write results back to a draft sheet keyed by SKU. Check what happens when you replay a row or receive two variants with the same name.
- Review the draft against the source row before importing it. For a Google Merchant Center feed, use its structured description attribute for AI-generated copy and follow the feed's digital source type requirements.
What to check
- Try the same item in a smaller variant and confirm it keeps its own dimensions and SKU.
- Reject unverified claims such as waterproofing or machine-washability before import. Check for stale descriptions when a source field changes.
- If you also send a product feed, confirm the approved description describes the same variant shown on the landing page.