Batch AI Generation

Generate AI-personalized content for hundreds or thousands of contacts at once. Learn how to use batch generation efficiently and monitor progress.

What is Batch Generation?

Instead of generating AI content for one contact at a time, batch generation processes multiple contacts simultaneously. This is the primary way to scale personalization in WarmOpener.

How to Use Batch Generation

Step 1: Prepare Your Contacts

Ensure your contacts have sufficient data for good personalization:

Basic fields: first_name, last_name, company, job_title
Optional: website, industry, recent_news, pain_point
More data = better AI output quality

Step 2: Select Contacts

On the Contacts page, you have multiple selection options:

Select All: Checkbox at the top selects all visible contacts
Individual Selection: Click checkboxes for specific contacts
Filter First: Use contact lists to narrow down before selecting

Step 3: Open AI Column Generator

With contacts selected, click Generate AI Column button
Modal opens with configuration options

Step 4: Configure Generation

Column Name: Name for the generated field (e.g., "Personalized_Opening", "Pain_Point"). Use underscores, no spaces.

Prompt: Describe what you want generated. Example: "Based on the contact's company and role, create a personalized opening sentence that references their industry challenges."

Website Enrichment: Toggle on to fetch company websites and extract deeper insights. Shows: X of Y contacts have website URLs.

Step 5: Review & Generate

Before generating, you'll see:

Contact count: Number of contacts that will be processed
Estimated cost: Approximate OpenAI API cost
Estimated time: How long the batch will take (~5-10 sec per contact)

Click Generate AI Content to start the batch.

Processing Behavior

Sequential Processing

Contacts are processed one at a time to avoid OpenAI rate limits. The modal shows:

Progress bar: X of Y contacts completed
Current contact being processed
Estimated time remaining

Error Handling

If a contact fails (e.g., API error):

System continues to next contact (doesn't stop entire batch)
Failed contacts are logged and shown in summary
You can retry failed contacts separately

Performance & Timing

Without Website Enrichment

Per contact: ~2-4 seconds
10 contacts: ~30 seconds
100 contacts: ~5 minutes
1,000 contacts: ~50 minutes

With Website Enrichment

Per contact: ~5-12 seconds (first time)
Cached: ~2-4 seconds (if website data already fetched)
100 contacts: ~10 minutes first time, ~5 minutes if cached

Best Practices

1. Test with Small Batch First

Before processing 1,000 contacts:

Select 5-10 representative contacts
Run generation and review output quality
Adjust prompt if needed
Then scale to full list

2. Use Clear, Specific Prompts

❌ Vague: "Write something personalized"

✅ Specific: "Create a 1-2 sentence opening that references their company's recent news or product launch, and explains how our solution addresses a specific challenge in their industry."

3. Batch Size Recommendations

Small batches (10-50): Good for testing and refinement
Medium batches (100-500): Standard workflow for campaigns
Large batches (1000+): Run during off-hours, expect 1+ hour processing

4. Re-Generation Strategy

If you need to regenerate for some contacts:

Old AI-generated data will be overwritten
Use same column name to replace existing data
Use different column name to create additional fields

Viewing Generated Data

After generation completes:

Generated field appears as a new column in Contacts table
Click contact row to expand and see full generated text
Use the field in email templates with placeholder syntax: {{Column_Name}}

⚠️ Keep Browser Tab Open

During batch generation, keep the browser tab open. Closing the tab or navigating away will cancel the in-progress batch. For large batches (500+ contacts), consider running during a time when you won't need to use your computer for other tasks.

💡 Pro Tip

Generate multiple AI columns for different parts of your email. For example: "Personalized_Opening", "Pain_Point", "Value_Prop", "CTA". This gives you modular components you can mix and match in different email templates.