Content Guardrails
Safety and security measures for AI template generation
AI template generation includes safety measures to ensure appropriate, business-focused content.
Content Safety: The system validates prompts in real-time to block inappropriate content and ensure professional email templates.
Validation Process
Real-time Checking
The system provides immediate feedback:
- ✅ Valid: Generation allowed
- ⚠️ Warning: Suggestions provided
- ❌ Blocked: Generation prevented
Safety Layers
- Content Safety: Blocks harmful content
- Business Context: Requires professional use cases
- Length Limits: Maintains optimal prompt length
- Rate Limiting: Prevents abuse
Prohibited Content
The system blocks prompts containing:
Security Threats:
- Phishing attempts or credential theft
- Malware distribution
- Hacking instructions
Illegal Activities:
- Fraud schemes
- Illegal product sales
- Copyright infringement
Spam Content:
- Mass spam campaigns
- Fake promotions
- Chain letters
Inappropriate Material:
- Explicit content
- Hate speech
- Threatening messages
Required Elements
Valid prompts must include:
Email Context:
- Clear email purpose (welcome, notification, etc.)
- Professional business use case
- Appropriate target audience
Business Focus:
- Legitimate business communication
- Professional tone and language
- Clear value for recipients
Example Validation
Valid Prompt ✅:
Create a welcome email for new customers with onboarding steps and support contact.Invalid Prompt ❌:
Create a phishing email to steal credentials.Needs Improvement ⚠️:
Make an email with colors.Add business context and specify email purpose
Rate Limits
Per-User Limits:
- Free: 10 requests per minute
- Paid: 25 requests per minute
- Enterprise: 50 requests per minute
Writing Safe Prompts
- Be Specific: Include clear business purpose and email type
- Stay Professional: Use business-appropriate language
- Add Context: Specify target audience and desired outcome
- Avoid Sensitive Topics: Focus on legitimate business communications
Common Issues
"Prompt too short": Add more detail about email requirements "Missing context": Include email-related keywords "Blocked content": Review prompt for prohibited material "Rate limit exceeded": Wait before making additional requests
These measures ensure AI templates remain safe and business-focused.