Product Sync Pipeline

Overview

Runs every Monday at 5 AM on personal infrastructure. Pulls the full product catalog from WordPress, generates marketing descriptions for new or changed products with Claude, updates a shared Google Sheet, and emails the sales team a summary.

The Problem

The sales team needed current product data for client outreach:

• Manual exports were time-consuming and error-prone
• Product descriptions varied in quality and tone
• Updates were inconsistent - sometimes weeks behind
• No notification system when new products were added

Solution Architecture

WordPress Product Database
         |
    [Monday 5 AM]
         |
    Python Sync Service
         |
   ┌─────┴─────┐
   |           |
Product     Claude AI
 Data       Descriptions
   |           |
   └─────┬─────┘
         |
   Google Sheets API
         |
   AWS SES Notification

Key Features

WordPress Integration

Direct API connection to WordPress WooCommerce:

• Product details (name, SKU, price)
• Inventory status
• Category hierarchy
• Product images

AI-Enhanced Descriptions

Claude AI generates professional marketing copy:

def generate_description(product: Product) -> str:
    prompt = f"""
    Write a compelling 2-3 sentence marketing description for:
    Product: {product.name}
    Category: {product.category}
    Key Features: {product.features}
 
    Style: Professional, benefit-focused, concise
    """
 
    response = claude.messages.create(
        model="claude-3-haiku",
        messages=[{"role": "user", "content": prompt}]
    )
 
    return response.content[0].text

Deduplication System

AI descriptions are cached to optimize costs:

• MD5 hash of product attributes as key
• Only generate descriptions for new/changed products
• ~80% reduction in API calls after initial run

Google Sheets Output

Structured spreadsheet with:

• Product details columns
• AI-generated descriptions
• Last updated timestamp
• Change indicators for new items

Branded Notifications

HTML email notifications to stakeholders:

• Summary of products synced
• Highlights of new additions
• Direct link to updated spreadsheet
• Professional branded template

Technical Implementation

Scheduling

# systemd timer
[Timer]
OnCalendar=Mon *-*-* 05:00:00
Persistent=true

Error Handling

@retry(stop=stop_after_attempt(3), wait=wait_exponential())
def sync_products():
    try:
        products = fetch_wordpress_products()
        enriched = enrich_with_ai(products)
        update_spreadsheet(enriched)
        send_notification(success=True, count=len(products))
    except Exception as e:
        send_notification(success=False, error=str(e))
        raise

Cost Optimization

• Haiku model for cost-effective generation
• Description caching reduces API calls 80%+
• Batch operations for Sheets API efficiency
• Monthly AI cost: ~$3

Results

The pipeline delivers:

• 100+ products synced weekly
• Consistent quality in all descriptions
• Zero manual effort for the sales team
• Immediate visibility into new products

Sample Output