# Prompt Optimization

Build Prompt Families that automatically select the best prompt for each input. Create specialized prompts that work together to perform better than any single prompt could.

## How Prompt Families Work

Instead of using one prompt for everything, Empromptu creates Prompt Families - collections of prompts that work together. Start with one prompt, and the system helps you build out the whole family.

### ❌Traditional Approach

Single Prompt → All Inputs → Inconsistent Results

### ✅Empromptu Approach

Input Analysis → Best Prompt from Family → Optimized Result

## 4 Optimization Interfaces

Access all interfaces through the Actions button on any task. Each tab handles different aspects of prompt optimization.

### 📊Event Log

Track every optimization attempt with detailed information:

**Timestamp:** When the optimization occurred  
**Input used:** What text was processed  
**Model:** Which AI model was used (e.g., gpt-4.1-mini)  
**Temperature:** Model settings (e.g., 0.2)  
**Response:** The generated output  
**Score:** Performance rating (0-10)  
**Score Reasoning:** Detailed explanation of why the score was assigned

### 👥Prompt Family

View and manage your collection of specialized prompts:

**Family Overview:** See all prompts working together  
**Individual Prompts:** View each family member with creation date  
**Status Control:** Activate/deactivate prompts  
**Performance Tracking:** See which prompts perform best

### 🔧Manual Optimization

Step-by-step improvement wizard for targeted enhancements:

**Step 1:** Select prompts to evaluate and optimize  
**Step 2:** Choose test inputs for scoring  
**Step 3:** Select evaluation criteria to optimize  
**Results:** Generate and test prompt variations

### ⚡Automatic Optimization

Let the system optimize for you with minimal input:

**Analysis:** System reviews current Prompt Family performance  
**Generation:** Creates new prompt variations automatically  
**Testing:** Runs new prompts against inputs and evaluations  
**Selection:** Adds best-performing prompts to your family

## Understanding Optimization Scores

### 0-10 Scoring Scale

9.0-10.0  
#### Excellent
Optimal performance, ready for business use

7.0-8.9  
#### Good
Business-ready, minor edge cases

5.0-6.9  
#### Acceptable
Improvable but functional

3.0-4.9  
#### Needs Attention
Moderate issues, requires optimization

0.0-2.9  
#### Poor
Requires immediate optimization

### Score Reasoning Examples

"extracted_completeness - AI response captures the essence and key emotional elements. Summary captures key experiential aspects while maintaining vivid language."

Score: 7.000

Each score includes detailed reasoning that explains which evaluations passed/failed, specific issues identified, and what contributed to the score.

## Optimization Strategies

### Start with Automatic

Best for establishing baseline performance and handling common optimization patterns.

1. Click "Start Automatic Optimization"  
2. System analyzes current performance  
3. New prompts generated and tested  
4. Results appear in Event Log  
5. Prompt Family grows automatically

### Refine with Manual

Best for addressing specific weaknesses and fine-tuning performance.

1. Review Event Log to identify problem areas  
2. Select specific inputs that performed poorly  
3. Choose evaluation criteria to focus on  
4. Run targeted optimization experiments  
5. Add successful variations to family

## Building Effective Prompt Families

### Include Different Scenarios

- Different input types (positive/negative content)  
- Different output requirements (brief/detailed)  
- Different edge cases (unusual inputs)

### Use Representative Test Data

- Test data should match real use cases  
- Include edge cases and challenging scenarios  
- Sufficient volume to identify patterns

### Monitor Performance Continuously

- Check Event Log regularly for new patterns  
- Update families as use cases evolve  
- Remove underperforming prompts

## Next Steps

1. Add better test data to improve your prompt optimization results  
2. Set up success criteria that guide prompt optimization effectively  
3. Find and fix problematic inputs using visual analysis

Get Started | View Documentation