From typing every product description to an agent writing your metafields

An agent writes product descriptions and fills metafields from your product data, cutting catalogue prep from hours to minutes per batch.

For anyone running a shopify catalogue · 8 steps · 7 min

How it works today.

You add a new product to Shopify and stare at the empty description box. You write something that sounds decent, hit save, then remember you need to fill the material metafield for filters, the care instructions for the product page app, and the SEO description that's slightly different from the main one. Each product means opening the same admin screens and typing variations of the same facts.

When you launch twenty products, that's twenty separate writing sessions. Your descriptions drift in tone because you wrote them on different days. Half your metafields stay empty because you forgot them or ran out of time. The product pages go live inconsistent and incomplete.

Before you start.

All of it has to be true, or step one fails in a way that is annoying to debug.

Shopify admin access with permission to edit products and metafields
A CSV export of your product data including titles, variants, prices, and any existing descriptions
Your brand voice guidelines or examples of descriptions you like
Metafield definitions already set up in Shopify for the fields you want filled

The steps.

Export your product data from Shopify
Go to Products in your Shopify admin, click Export, and select 'Current page' or 'All products'. Choose CSV format and include all product information. Download the file and open it to confirm you have columns for handle, title, body HTML, vendor, product type, and any existing metafield values.
Map your metafields to the agent's output format
List every metafield you want filled: material, care instructions, key features, SEO description, or others. Note the exact metafield namespace and key from your Shopify metafield definitions. The agent needs to know these exact names to format its output correctly.

Set up the agent with your product writing requirements

Configure your agent with access to process CSV files and generate structured text output. Test that it can read your product export and return formatted responses. If using a no-code builder, confirm it has a CSV import step and text generation capabilities.

Paste this

You write Shopify product descriptions and metafields. I will give you product data as CSV rows. For each product, write: 1) A product description in [your brand voice - conversational/technical/luxury], 2) Fill these metafields with the format namespace.key: value. My metafields are: [list your exact metafield namespaces and keys]. Keep descriptions under 200 words. Use the product title, type, and vendor to infer details when needed.

Process a test batch of five products
Copy five rows from your CSV and paste them into the agent. Review the generated descriptions for tone and accuracy. Check that metafield outputs match your namespace.key format exactly. If descriptions are too generic or metafields are wrong, refine your prompt and test again.
Run the full product batch through the agent
Feed your complete product CSV to the agent in manageable chunks of 10-20 products. The agent will return descriptions and metafield values for each product. Save this output as a new CSV or structured format that you can work with for the import step.
06
Format the agent output for Shopify import
Create a new CSV with columns for Handle, Body HTML, and each metafield column using Shopify's import format. Copy the agent's descriptions into Body HTML and metafield values into their respective columns. Ensure handles match exactly between your original export and this import file.
Paste this
```
Convert this product data into Shopify CSV import format. Create columns for: Handle, Body HTML (for description), [your metafield column names]. Format metafields as single values, not namespace.key format. Keep the exact product handles from the original data.
```
Import the updated descriptions and metafields
In Shopify admin, go to Products > Import and upload your formatted CSV. Select 'Update existing products' and map your CSV columns to the correct Shopify fields. Run the import and check the preview before confirming. Shopify will update only the fields you included.
Review imported products for accuracy and consistency
Spot-check 10-15 products from different categories. Verify descriptions match your brand voice and metafields populated correctly. Look for any formatting issues or missing data. Fix any problems manually and note them for improving your agent prompt next time.