Products Training Context

Import and manage product data to help your AI assistant answer questions about your products, pricing, and availability.

Overview

Product data enables your AI to:

• Answer questions about product features
• Provide accurate pricing information
• Explain product specifications
• Recommend products based on customer needs
• Check product availability

Import Methods

CSV Upload

Upload a CSV file directly from your computer:

1. Click "Import CSV" button
2. Select your CSV file
3. Add optional context about the data
4. Click "Import"

File Requirements:

• Format: CSV (comma-separated values)
• Max file size: 10MB
• Required columns: name
• Optional columns: price, category, and custom columns

CSV URL

Import products from a remote CSV URL:

1. Click "Import CSV"
2. Select "Remote URL" option
3. Enter CSV URL
4. Add optional context
5. Click "Import"

URL Requirements:

• Must be publicly accessible
• Must return valid CSV content
• Supports HTTP and HTTPS

API Integration

Connect to an external API to sync products automatically:

1. Configure API endpoint
2. Set sync interval
3. Enable auto-rescan
4. Products sync automatically

API Requirements:

• Must return JSON array of products
• Must include product name
• Can include custom fields

CSV Format

Required Columns

• name: Product name (required)

Optional Columns

• price: Product price (numeric)
• category: Product category (text)
• description: Product description (text)
• currency: Price currency (default: USD)
• Custom columns: Any additional fields you need

Example CSV

name,price,category,description
Widget A,29.99,Electronics,High-quality widget with advanced features
Widget B,19.99,Electronics,Budget-friendly widget for everyday use
Service Package,199.99,Services,Complete service package with support

Column Mapping

If your CSV uses different column names, you can map them:

• Map product_name → name
• Map cost → price
• Map type → category

Feed Context

The "Context" field (optional) provides additional information about your product data:

Example contexts:

• "These are cars available for sale"
• "These are software products we offer"
• "These are service packages with pricing"

Why it helps:

• Helps AI understand the data better
• Improves response accuracy
• Enables better product recommendations

Auto-Rescan Configuration

For CSV URL and API imports, you can configure automatic re-scanning:

Rescan Intervals

• Hourly: Rescans every hour
• Every 6 Hours: Rescans 4 times per day
• Every 12 Hours: Rescans twice per day
• Daily: Rescans once per day
• Weekly: Rescans once per week

How It Works

1. System checks external URL/API at set interval
2. Compares new data with existing products
3. Updates changed products
4. Adds new products
5. Removes products no longer in feed (optional)

Sync Status

Monitor sync status:

• Last synced: When data was last updated
• Next sync: When next automatic sync will occur
• Sync status: Success, error, or pending
• Error messages: If sync fails, error details shown

Product Management

Viewing Products

View all imported products in a table showing:

• Product name
• Category
• Price (with currency)
• Character count

Character Counting

Each product's character count includes:

• Product name
• Description
• Category
• All custom field values

Total character count affects your usage limit.

Updating Products

Products can be updated by:

• Re-importing CSV with updated data
• Updating via API (if configured)
• Automatic sync (if auto-rescan enabled)

Background Processing

After products are imported or updated, they are processed in the background:

• Chunking: Product data is split into manageable chunks
• Embedding Generation: AI embeddings are created for each chunk
• Storage: Chunks and embeddings are stored for retrieval

Processing happens asynchronously:

• Products are imported immediately
• Chunking and embedding happen in background jobs
• Products become available to AI as processing completes
• You can monitor job status via the Job Queue API

Typical processing times:

• Small products: 30 seconds - 1 minute per product
• Medium products: 1-2 minutes per product
• Large products (with long descriptions): 2-5 minutes per product

Monitoring Processing

Check Job Status:

• Use the Job Queue API to check job status
• View processing progress and completion
• Identify any failed jobs that need attention

Processing Indicators:

• Pending: Job is queued, waiting to start
• Processing: Job is actively being processed
• Completed: Product is ready for AI use
• Failed: Processing failed (check error details)

Best Practices

1. Keep Data Current: Update product data regularly
2. Use Clear Names: Product names should be descriptive
3. Include Descriptions: Detailed descriptions help AI provide better answers
4. Organize by Category: Use categories to group related products
5. Monitor Sync Status: Check sync logs regularly for errors
6. Optimize Character Usage: Remove unnecessary data to stay within limits

Troubleshooting

Import Fails

• Check CSV format is valid
• Verify required columns are present
• Ensure file size is under 10MB
• Check URL is accessible (for URL imports)

Sync Errors

• Verify API endpoint is accessible
• Check API returns valid JSON
• Review error messages in sync status
• Test API endpoint manually

Character Limit Exceeded

• Delete unused products
• Remove unnecessary descriptions
• Optimize product data
• Upgrade plan for higher limits

Next Steps

• View Overview for general training context information
• Learn about Documents for document management
• Learn about Website for website crawling
• Learn about Instructions for custom instructions
• Monitor processing with Job Queue API