Prompt Engineering Mastery

123456789101112
{
  "system": "You are a senior financial analyst specializing in SaaS metrics. You communicate complex financial concepts clearly and always support conclusions with specific data. When uncertain, you acknowledge limitations rather than speculating.",
  "context": {
    "company_stage": "Series B",
    "industry": "B2B SaaS",
    "arr": "$12M",
    "growth_rate": "120% YoY"
  },
  "task": "Analyze the provided financial metrics and identify the top 3 areas of concern for the upcoming board meeting.",
  "output_format": {
    "structure": "numbered_list",
    "per_item": ["concern_title", "supporting_data", "recommended_action"],

123456789101112
{
  "system": "You are a customer support ticket classifier. Classify tickets into exactly one category: billing, technical, feature_request, or general. Respond with only the category name.",
  "examples": [
    {
      "user": "I was charged twice for my subscription this month",
      "assistant": "billing"
    },
    {
      "user": "The app crashes when I try to export to PDF",
      "assistant": "technical"
    },
    {

123456789101112
const systemPrompt = `
# Identity
You are ${config.assistantName}, an AI assistant for ${config.companyName}. You specialize in ${config.expertise} and help users with ${config.primaryUseCases}.

# Behavioral Guidelines
- Always be helpful, accurate, and respectful
- Acknowledge uncertainty rather than guessing
- ${config.customGuidelines.join('\n- ')}

# Response Format
- Keep responses concise unless detail is requested
- Use markdown formatting for readability

12345678910
{
  "system": "You are a senior product analyst specializing in voice-of-customer analysis. You identify patterns, prioritize by business impact, and provide actionable recommendations. Always cite specific feedback quotes to support conclusions.",
  
  "user_prompt": "Analyze these customer feedback items and provide:\n1. Top 3 themes with frequency count\n2. Sentiment breakdown (positive/neutral/negative)\n3. Priority ranking based on: revenue impact, user volume affected, implementation effort\n4. Specific product recommendations\n\nFeedback items:\n{feedback_array}\n\nOutput as structured JSON with the following schema:\n{\n  \"themes\": [{\"name\": string, \"count\": number, \"sample_quotes\": string[]}],\n  \"sentiment\": {\"positive\": number, \"neutral\": number, \"negative\": number},\n  \"priorities\": [{\"issue\": string, \"score\": 1-10, \"rationale\": string}],\n  \"recommendations\": [{\"action\": string, \"expected_impact\": string, \"effort\": \"low|medium|high\"}]\n}",
  
  "few_shot_example": {
    "input": "Sample of 50 feedback items about checkout flow...",
    "output": "{ themes: [{name: 'Payment failures', count: 23, ...}], ... }"
  }
}

123456789101112
import anthropic

def analyze_with_validation(user_request: str) -> dict:
    client = anthropic.Anthropic()
    
    # Step 1: Classify and extract
    classification = client.messages.create(
        model="claude-sonnet-4-20250514",
        max_tokens=500,
        system="Classify the request type and extract key entities. Output JSON only.",
        messages=[{"role": "user", "content": user_request}]
    )

123456789101112
interface PromptVersion {
  id: string;
  version: string;
  systemPrompt: string;
  userTemplate: string;
  trafficWeight: number;
  createdAt: Date;
  metrics: {
    totalRequests: number;
    avgLatency: number;
    successRate: number;
    userSatisfaction: number;