E-commerce product description and image enhancement from supplier data

Introduction

Running an e-commerce business means you're constantly wrestling with supplier data that arrives in various formats, often incomplete or poorly formatted. A spreadsheet of product names comes in, but the descriptions read like technical specs written by engineers. Images exist but they're undersized, poorly lit, or missing entirely. Your customers deserve better, but manually rewriting hundreds of descriptions and enhancing images would consume weeks of work.

This is where combining multiple AI tools into an automated workflow becomes genuinely valuable. Instead of manually copying supplier data into different tools and stitching results together, you can set up a single pipeline: data enters at one end, passes through enhancement stages, and emerges ready for your storefront. No human intervention required between steps.

In this Alchemy recipe, we'll combine three specialist tools: ai-boost for analysing and enhancing your raw product data; copy-ai for generating polished, conversion-focused product descriptions; and nsketch-ai for upscaling and improving your product images. We'll wire them together using an orchestration platform so that when new supplier data arrives, the entire enhancement process happens automatically.

The Automated Workflow

Choosing Your Orchestration Platform

For this workflow, I'd recommend starting with Zapier if you want the quickest setup with minimal technical friction. However, if you need finer control over data transformation or plan to run this frequently, n8n offers better value and more flexibility. Make (formerly Integromat) sits in the middle, offering decent conditional logic without the learning curve of n8n.

For this guide, I'll show examples using n8n, which gives you the most control whilst remaining accessible. The logic translates easily to Zapier or Make if you prefer those platforms.

Architecture Overview

Your workflow follows this sequence:

1. New supplier data arrives (CSV upload, API call, or database record)
2. ai-boost analyses the data and extracts key product information
3. copy-ai generates a compelling product description using the extracted data
4. nsketch-ai upscales and enhances the product image
5. Enhanced data is saved to your e-commerce system

This happens entirely automatically once you've set it up. Supplier data goes in; polished products come out.

Step 1: Receiving Supplier Data

Start by defining how data enters your workflow. Most suppliers deliver data via CSV file, API endpoint, or database export. We'll use a webhook trigger to keep things flexible.

Copy
POST /webhook/product-upload
Content-Type: application/json

{
  "product_id": "SUPP-12345",
  "supplier_name": "TechParts Ltd",
  "product_title": "USB-C Cable 2M",
  "raw_description": "High quality cable connector type USB-C length 2 meters for charging and data transfer",
  "image_url": "https://supplier.example.com/images/usb-cable-12345.jpg",
  "category": "Cables & Connectors",
  "supplier_sku": "TCP-2M-001",
  "price": 12.50
}

In n8n, create a Webhook trigger node and set it to POST. This gives you an endpoint where you can send data. If your supplier delivers files instead, use the File trigger node to watch a folder.

Step 2: Analyse with ai-boost

ai-boost's core strength is extracting structured information from unstructured data. It understands product context and can categorise, tag, and identify missing information.

Create an HTTP Request node in n8n to call the ai-boost API:

Copy
POST https://api.ai-boost.com/v1/analyse
Content-Type: application/json
Authorization: Bearer YOUR_AIBOOST_API_KEY

{
  "content": "High quality cable connector type USB-C length 2 meters for charging and data transfer",
  "product_title": "USB-C Cable 2M",
  "analysis_type": "product_enhancement",
  "extract_fields": [
    "material",
    "connector_type",
    "length",
    "use_cases",
    "key_features",
    "compatibility",
    "warranty_info"
  ]
}

ai-boost returns structured data that identifies missing information and enriches what's present:

Copy{
  "extracted_data": {
    "connector_type": "USB-C",
    "length": "2 metres",
    "use_cases": ["charging", "data transfer"],
    "key_features": ["high quality", "durable"],
    "material": "null",
    "compatibility": ["USB-C devices"],
    "warranty_info": "null"
  },
  "quality_score": 0.62,
  "missing_fields": ["material", "warranty_info", "cable_rating"],
  "enrichment_suggestions": [
    "Specify copper construction for conductivity",
    "Add warranty period",
    "Mention compatibility with popular devices"
  ]
}

In your n8n workflow, map this output to a new variable. You'll use this structured data in the next step.

Step 3: Generate Description with copy-ai

Now that you have clean, structured product information, copy-ai can write a proper description. Rather than improving the original supplier text, you're giving it structured inputs so it writes from scratch.

Create another HTTP Request node:

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

{
  "template": "ecommerce_product_description",
  "tone": "professional_friendly",
  "inputs": {
    "product_name": "USB-C Cable 2M",
    "key_features": ["USB-C connector", "2 metre length", "charging and data transfer"],
    "use_cases": "charging smartphones, tablets, and laptops; syncing data",
    "compatibility": "all USB-C compatible devices",
    "enrichment": "Heavy-duty construction for durability"
  },
  "max_length": 200,
  "style": "conversion_focused"
}

copy-ai returns a polished description:

Copy{
  "description": "Reliable 2-metre USB-C cable for fast charging and seamless data transfer across all your USB-C devices. Built for durability with heavy-duty construction, this cable handles daily use without wear. Perfect for smartphones, tablets, and laptops.",
  "variants": [
    {
      "tone": "technical",
      "text": "Universal USB-C cable featuring full-speed data transfer at USB 3.0 specifications..."
    }
  ]
}

Store this description in a variable for later use.

Step 4: Enhance Image with nsketch-ai

Meanwhile, you'll process the product image. nsketch-ai specialises in upscaling and enhancing product photos, making them look professional even if the original is mediocre.

Create an HTTP Request node that calls nsketch-ai:

Copy
POST https://api.nsketch-ai.com/v1/enhance
Content-Type: application/json
Authorization: Bearer YOUR_NSKETCH_API_KEY

{
  "image_url": "https://supplier.example.com/images/usb-cable-12345.jpg",
  "enhancement_type": "product_photo",
  "upscale_factor": 2,
  "adjustments": {
    "brightness": "auto",
    "contrast": "enhance",
    "colour_correction": true,
    "background_removal": false,
    "shadow_fill": "light"
  }
}

nsketch-ai returns a URL to the enhanced image:

Copy{
  "original_url": "https://supplier.example.com/images/usb-cable-12345.jpg",
  "enhanced_url": "https://cdn.nsketch-ai.com/enhanced/usb-cable-12345-enhanced.jpg",
  "enhancement_applied": {
    "upscale": "2x",
    "brightness_adjustment": "+15%",
    "contrast_adjustment": "+20%"
  },
  "file_size": "2.4MB"
}

The workflow can now download this enhanced image and save it to your own storage (AWS S3, your server, etc.) so you own the file rather than relying on nsketch-ai's CDN.

Step 5: Save to Your E-commerce System

Finally, combine all the enhanced data and push it to your e-commerce platform. Most platforms offer an API for bulk product updates.

Copy
POST https://yourstorefront.example.com/api/v1/products
Content-Type: application/json
Authorization: Bearer STORE_API_KEY

{
  "product_id": "SUPP-12345",
  "title": "USB-C Cable 2M",
  "description": "Reliable 2-metre USB-C cable for fast charging and seamless data transfer across all your USB-C devices...",
  "image_url": "https://your-cdn.com/products/usb-cable-12345-enhanced.jpg",
  "category": "Cables & Connectors",
  "sku": "TCP-2M-001",
  "price": 12.50,
  "tags": ["USB-C", "charging cable", "data transfer"],
  "metadata": {
    "source": "TechParts Ltd",
    "enhancement_date": "2025-01-15T10:30:00Z",
    "quality_score": 0.62
  }
}

In n8n, add a final HTTP Request node pointing to your store's API. Map all the variables from previous steps into this request body.

Complete n8n Workflow Structure

Here's how your nodes connect:

1. Webhook (trigger): Receives incoming supplier data
2. Set Variable: Store the incoming payload
3. HTTP Request: Call ai-boost API
4. Set Variable: Store extracted data
5. HTTP Request: Call copy-ai API
6. Set Variable: Store generated description
7. HTTP Request: Call nsketch-ai API
8. HTTP Request: Download enhanced image to your storage
9. HTTP Request: POST to your store's product API
10. Send Slack notification (optional): Confirm completion

Use n8n's built-in error handling on each HTTP node. If ai-boost fails, you can route to a fallback (use the original description) or pause and notify you.

Rate Limiting Considerations

Each API has rate limits. ai-boost typically allows 100 requests per minute on their standard plan; copy-ai allows 50 per minute; nsketch-ai allows 30 per minute. If you're processing thousands of products, add delay nodes between requests or batch your uploads.

In n8n, use the Wait node to add deliberate pauses:

Copy
{
  "type": "wait",
  "duration": 2,
  "unit": "seconds"
}

The Manual Alternative

If you prefer not to set up automation initially, you can run this workflow semi-manually using individual tool dashboards.

Step 1: Upload your CSV to ai-boost's web interface and download the enriched output.

Step 2: Copy the enriched data into copy-ai's editor, select your template, and generate descriptions one batch at a time.

Step 3: Upload images to nsketch-ai's batch processor and download the enhanced versions.

Step 4: Manually import the enhanced descriptions and images into your e-commerce admin panel.

This works fine for small batches (50 products or fewer), but the manual handoffs become tedious quickly. Once you automate, you'll realise how much time you were losing to copying and pasting between tools.

Pro Tips

Error Handling and Retry Logic

Set up retry logic on your HTTP requests. If copy-ai's API is temporarily unavailable, you want the workflow to retry after 30 seconds rather than failing immediately. In n8n, use the "Retry" option on each HTTP node: set it to retry 3 times with exponential backoff.

Copy
{
  "retryConfig": {
    "maxRetries": 3,
    "initialDelay": 1000,
    "backoffMultiplier": 2
  }
}

Track Processing Status

Add a status field to your product records so you can track which ones have been enhanced. Mark products as "pending", "processing", "completed", or "failed". This helps you reprocess failed items without re-running successful ones.

Batch Processing vs Real-time

If suppliers send 500 products at once, processing them all in parallel might hit rate limits. Instead, add a "Process in batches of 10" node. n8n's "Loop" functionality lets you process items sequentially with built-in delays.

Monitor Costs

Each API call costs money. ai-boost typically costs $0.001 per enrichment; copy-ai costs $0.002 per description; nsketch-ai costs $0.005 per image. For 1,000 products, that's $8 total in API costs, far cheaper than hiring someone to do this work manually. But monitor your actual usage through each platform's dashboard to avoid surprises.

Store Original Data

Always keep the original supplier data alongside the enhanced version. This lets you audit what changed and reprocess if you adjust your ai-boost or copy-ai settings later. Store originals in a separate database table or archive folder.

Cost Breakdown

For a small e-commerce business processing 500 products per month, your total cost is approximately £247 (using n8n self-hosted). If you process 2,000 products monthly, you'll hit overage charges on ai-boost and nsketch-ai, bringing costs to around £320-350. Still cheaper than a part-time employee handling this manually.