E-commerce product description and image enhancement from supplier data

You're sat in your e-commerce operations meeting. Another supplier has sent over 500 product listings in a spreadsheet. Generic descriptions, inconsistent naming, no images. Your team groans. Someone says, "We'll have to rewrite these manually," and you know that means three days of grinding through copy, hunting down stock photos, and reformatting everything to match your brand guidelines.............

This doesn't need to happen anymore. If you have supplier data (product names, specs, basic attributes), you can automatically generate polished product descriptions and commissioned artwork without touching a single spreadsheet yourself. The workflow runs in the background; your products arrive in your system ready to sell.

This guide shows you how to build that exact pipeline using three focused tools: ai-boost for image generation, copy-ai for description writing, and nsketch-ai for additional visual assets. We'll wire them together using one of four orchestration platforms so that when supplier data arrives, everything else happens automatically.

The Automated Workflow

Setting Up Your Data Source

Your workflow needs a trigger. In practice, this is usually one of the following: a new file uploaded to cloud storage, a new row added to a Google Sheet, data arriving via webhook, or a scheduled check of your supplier's API.

For this example, we'll use Zapier as our orchestration platform, though we'll mention n8n alternatives where relevant. Zapier offers the simplest setup for beginners, but if you're comfortable with self-hosted solutions or need more complex logic, n8n or Make work equally well.

The trigger point matters. If your supplier delivers files weekly, set Zapier to check a shared Google Drive folder every Monday morning. If they use an API, connect directly to that endpoint. If they email CSV files, use an email parser.

Here's a basic webhook structure for accepting supplier data directly:

Copy
POST /api/v1/supplier-products
Content-Type: application/json

{
  "supplier_id": "SUP-12345",
  "products": [
    {
      "sku": "CHAIR-001",
      "name": "Wooden Office Chair",
      "category": "Furniture",
      "material": "Oak",
      "dimensions": "45cm W x 80cm H x 50cm D",
      "price_wholesale": 45.00,
      "keywords": "ergonomic, office, wood"
    }
  ]
}

When this data arrives, Zapier (or n8n, or Make) will parse it and pass each product through three sequential steps.

Step 1: Generate Product Descriptions with Copy-AI

Copy-ai specialises in writing product copy from raw product data. The tool accepts product attributes and returns marketing-ready descriptions tailored to your brand voice.

Set up a Copy-AI integration in Zapier like this:

1. In Zapier, add an action step: "Copy-AI — Generate Product Description"
2. Map your incoming fields to Copy-AI's input schema
3. Specify your brand voice in the prompt template

Here's what the request payload looks like:

Copy
POST https://api.copy-ai.com/v1/generate-description
Authorization: Bearer YOUR_COPY_AI_API_KEY
Content-Type: application/json

{
  "product_name": "Wooden Office Chair",
  "product_attributes": {
    "material": "Oak",
    "dimensions": "45cm W x 80cm H x 50cm D",
    "category": "Furniture"
  },
  "keywords": "ergonomic, office, wood",
  "brand_voice": "Professional, concise, emphasise sustainability",
  "tone": "informative",
  "length": "medium",
  "include_specs": true
}

Copy-AI returns a complete product description:

Copy{
  "description": "Hand-crafted oak office chair combining timeless design with modern ergonomic support. Built from sustainably sourced oak with a contoured seat and adjustable height mechanism. Dimensions: 45cm W x 80cm H x 50cm D. Perfect for professionals who demand both style and comfort.",
  "meta_description": "Premium oak office chair with ergonomic design",
  "keywords_used": ["ergonomic", "office", "wood", "sustainable"],
  "generation_time_ms": 340
}

Copy the description field and store it in your workflow state. You'll reference it later when uploading to your e-commerce platform.

Critical point here: Copy-AI respects rate limits. Free plans allow 10 requests per minute; paid tiers extend this. If you're processing 500 products, stagger requests or use a batch endpoint to avoid throttling.

Step 2: Commission Images with AI-Boost

Now your product has a description. Next, you need a visual. AI-boost generates product photography from product attributes, supplier images, or text descriptions. It's particularly useful if your supplier hasn't provided images, or if their images are poor quality. For more on this, see Product photography studio with AI model training and mar....

In Zapier, add a second action: "AI-Boost — Generate Product Image"

The setup:

Copy
POST https://api.ai-boost.io/v1/images/generate
Authorization: Bearer YOUR_AI_BOOST_API_KEY
Content-Type: application/json

{
  "product_name": "Wooden Office Chair",
  "product_description": "Hand-crafted oak office chair combining timeless design with modern ergonomic support. Built from sustainably sourced oak with a contoured seat and adjustable height mechanism.",
  "category": "Furniture",
  "image_style": "professional-product-photography",
  "background": "white",
  "dimensions": "1200x1200",
  "quantity": 1
}

AI-Boost generates and returns URLs to the generated images:

Copy{
  "status": "success",
  "images": [
    {
      "url": "https://images.ai-boost.io/generated/12345-chair-001.png",
      "width": 1200,
      "height": 1200,
      "file_size_kb": 450,
      "style_applied": "professional-product-photography"
    }
  ],
  "generation_time_seconds": 12,
  "credits_used": 2
}

Store this image URL in your workflow. You'll need it when syncing to your e-commerce system.

Note on costs: AI-Boost charges per image generated. A typical product photo costs between 1-3 credits, and credits range from £0.50 to £2.00 each depending on your subscription tier. Factor this into your monthly budget when processing large batches.

Step 3: Create Additional Assets with nSketch-AI

nSketch-AI specialises in sketch-style and illustrative assets. Use it to generate lifestyle images, infographics, or alternative visual representations of your products. This is optional for basic workflows, but valuable if you want multiple images per product or lifestyle photography to accompany your product photography.

In Zapier, add a third action: "nSketch-AI — Generate Lifestyle Image"

Copy
POST https://api.nsketch.io/v1/generate-lifestyle
Authorization: Bearer YOUR_NSKETCH_API_KEY
Content-Type: application/json

{
  "product_name": "Wooden Office Chair",
  "context": "modern office environment",
  "style": "realistic-lifestyle",
  "include_product": true,
  "environment": "home-office",
  "mood": "professional, inviting"
}

nSketch returns a lifestyle image showing your product in context:

Copy{
  "status": "success",
  "image": {
    "url": "https://images.nsketch.io/generated/67890-chair-lifestyle.png",
    "width": 1024,
    "height": 768,
    "style_applied": "realistic-lifestyle"
  },
  "generation_time_seconds": 18
}

Store this URL separately. Most e-commerce platforms allow multiple images per product, so assign this as your secondary lifestyle image.

Step 4: Sync Everything to Your E-Commerce Platform

Now you have the complete product package: description, hero image, and lifestyle shot. The final step in your workflow is pushing this data to your actual e-commerce system.

If you use Shopify, WooCommerce, or any major platform, Zapier has native integrations. Here's what a Shopify sync looks like:

Copy
POST https://your-store.myshopify.com/admin/api/2024-01/products.json
Authorization: Bearer YOUR_SHOPIFY_API_KEY
Content-Type: application/json

{
  "product": {
    "title": "Wooden Office Chair",
    "body_html": "<p>Hand-crafted oak office chair combining timeless design with modern ergonomic support. Built from sustainably sourced oak with a contoured seat and adjustable height mechanism. Dimensions: 45cm W x 80cm H x 50cm D. Perfect for professionals who demand both style and comfort.</p>",
    "vendor": "Supplier Name",
    "product_type": "Furniture",
    "tags": "ergonomic, office, wood",
    "images": [
      {
        "src": "https://images.ai-boost.io/generated/12345-chair-001.png",
        "alt": "Wooden Office Chair product photography"
      },
      {
        "src": "https://images.nsketch.io/generated/67890-chair-lifestyle.png",
        "alt": "Wooden Office Chair in office environment"
      }
    ]
  }
}

Shopify returns a product object with IDs. Store these IDs if you need to reference products later (for inventory updates, reviews, etc.).

Putting It All Together in Zapier

Your complete Zapier workflow looks like this:

1. Trigger: New row in Google Sheet (or webhook, or email attachment parsed)
2. Action 1: Copy-AI — Generate description
3. Action 2: AI-Boost — Generate product image
4. Action 3: nSketch-AI — Generate lifestyle image
5. Action 4: Shopify (or your platform) — Create product

Between each step, use Zapier's "Formatter" tool to restructure data if needed. For instance, if Copy-AI returns a JSON object but Shopify expects plain text, add a formatter step.

Example formatter setup:

Copy
Input:
{
  "description": "Hand-crafted oak office chair...",
  "meta_description": "Premium oak office chair...",
  "keywords_used": ["ergonomic", "office", "wood"]
}

Transform using:
Text — Remove HTML tags (none here, but good practice)
Text — Trim whitespace

Output:
"Hand-crafted oak office chair combining timeless design with modern ergonomic support..."

If you're using n8n instead of Zapier, the workflow structure is identical, but you'll write JSON directly. Here's the n8n equivalent:

Copy{
  "nodes": [
    {
      "name": "Trigger",
      "type": "n8n-nodes-base.webhook",
      "position": [100, 200]
    },
    {
      "name": "Copy-AI",
      "type": "n8n-nodes-base.httpRequest",
      "position": [300, 200],
      "parameters": {
        "url": "https://api.copy-ai.com/v1/generate-description",
        "method": "POST",
        "headers": {
          "Authorization": "Bearer YOUR_COPY_AI_API_KEY"
        },
        "body": "=JSON.stringify({product_name: $json.name, product_attributes: $json.attributes})"
      }
    }
  ],
  "connections": {
    "Trigger": {
      "main": [
        [
          {
            "node": "Copy-AI",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  }
}

The Manual Alternative

If you want human review before products go live, inject a manual approval step. In Zapier, add a "Send Email Digest" action before the final sync. Email yourself the generated description and images, review them for 2 minutes, then approve or reject.

This works well if your supplier data is often incomplete or contradictory. AI tools make reasonable guesses, but human eyes catch mistakes (wrong colours, missing material information, pricing errors) before they reach your customers.

Alternatively, use a "Pause" action in Zapier to hold each product in a queue, then manually push them through Slack or a simple approval interface.

For higher-volume operations, this manual step becomes a bottleneck. Test with 10 products first, measure how many need correction, then decide whether full automation or hybrid review makes sense for your business.

Pro Tips

Rate Limiting and Batching

If you're processing 500 products, you'll hit API rate limits. Copy-AI allows 10 requests per minute on free plans. Instead of triggering all 500 at once, use Zapier's "Delay" action to space requests out: 6 seconds between each. This keeps you under rate limits and prevents API rejections.

Alternatively, ask Copy-AI about their batch endpoint, which processes multiple products in a single request at higher throughput.

Image Storage and CDN

AI-Boost and nSketch-AI return temporary URLs. Download these images and store them on your own server or a CDN (Cloudflare, AWS S3, Bunny CDN). This gives you permanent access and faster loading times. Many e-commerce platforms allow you to upload images directly rather than referencing external URLs; do this when possible.

Error Handling and Retry Logic

Networks fail. APIs time out. Set up error handling in Zapier by adding "Only Continue If" conditions and retry logic.

For example:

Copy
If Copy-AI request fails:
  - Wait 30 seconds
  - Retry up to 3 times
  - If still failing, send alert email and pause workflow

Zapier's "Paths" feature handles this. Create two branches from the Copy-AI action: one for success, one for failure.

Customising Brand Voice Per Platform

If you sell on multiple platforms (Shopify, Amazon, eBay), use different prompts for each. Copy-AI descriptions for Amazon should emphasise bullet-point specs; Shopify descriptions can be more narrative. Store these templates as Zapier variables and swap them based on your trigger.

Cost Optimisation

Run your workflow during off-peak hours. If Copy-AI or AI-Boost charge less during nights and weekends, schedule your supplier batch imports for those times. Monitor your API usage monthly; if you're consistently using only 30% of your subscription tier, downgrade.

Track the cost per product:

(Copy-AI cost + AI-Boost cost + nSketch-AI cost + Zapier cost) / number of products processed

If this exceeds your manual labour cost, consider a hybrid approach: use AI for descriptions only, hire interns for image sourcing.

Cost Breakdown

The cost per product ranges from £1.67 (small batch) to £0.60 (large batch). Compare this to paying someone £15 per hour to write descriptions and find images. At that rate, manual work costs £5-10 per product. Automation breaks even after 100 products and saves money thereafter.