Step-by-Step Template Creation
1
Navigate to Templates
Click Templates in the sidebar
2
Click Create Template
Click the Create Template button
3
Fill in Basic Details
- Name: Descriptive identifier (e.g., “SaaS Pain Point Hook”)
- Category: Select or type custom
- Output Column Name: What the result column will be called
4
Write Your Prompt
Enter your AI instructions with
{{variables}}5
Configure Constraints
- Character limit: Max characters (optional)
- Word limit: Max words (optional)
- Forbidden words: Words to never include (optional)
6
Set AI Parameters
- Model: gpt-4o-mini (default, recommended)
- Temperature: 0.7 (default, 0 = focused, 1 = creative)
- Max Tokens: 150 (default output length)
7
Save Template
Click Create Template to save
Writing Effective Prompts
Be specific about output format
Be specific about output format
❌ “Write something about their company”✅ “Write a 1-sentence opening hook mentioning their company’s recent growth”
Include what NOT to include
Include what NOT to include
Specify length clearly
Specify length clearly
Give examples if helpful
Give examples if helpful
Reference your actual offer
Reference your actual offer
Output Column Name
The Output Column Name determines:- What the new CSV column will be called
- How you reference it in template chaining
- What variable name appears in your sending tool
Character and Word Limits
Set constraints to control output length:| Setting | Effect |
|---|---|
| Character Limit | AI tries to stay under X characters |
| Word Limit | AI tries to stay under X words |
| Both | AI respects whichever is more restrictive |
Limits are guidelines for the AI, not hard cutoffs. Outputs may occasionally exceed limits slightly. The quality flag system will catch these.
Forbidden Words
Add words that should never appear in outputs:- Enter as comma-separated list or tags
- Common uses: competitor names, banned phrases, overused words
- Outputs containing forbidden words get flagged
AI Model Settings
| Setting | Default | Recommendation |
|---|---|---|
| Model | gpt-4o-mini | Keep default for best cost/quality |
| Temperature | 0.7 | Lower (0.3) for consistent, higher (0.9) for creative |
| Max Tokens | 150 | Increase for longer outputs |
Testing Before Use
After creating a template, test it before running a full job:- Open the template
- Select a file to test with
- Click Test
- Review outputs and quality flags
- Iterate on the prompt if needed