Analyze Sentiment and Feedback

Analyze the sentiment and extract actionable insights from the following feedback.

Feedback Text:
{{feedback_text}}

Source Type:
{{source_type}}

Return your analysis as JSON:

{
  "overall_sentiment": "[positive/negative/neutral/mixed]",
  "sentiment_score": [-1.0 to 1.0],
  "emotions_detected": ["[emotion1]", "[emotion2]"],
  "key_themes": [
    {
      "theme": "[Topic or aspect mentioned]",
      "sentiment": "[positive/negative/neutral]",
      "verbatim": "[Exact quote from feedback]"
    }
  ],
  "praise_points": ["[What they liked]"],
  "pain_points": ["[What they disliked or struggled with]"],
  "suggestions": ["[Any explicit or implicit suggestions]"],
  "urgency": "[high/medium/low/none]",
  "churn_risk": "[high/medium/low]",
  "response_recommended": true/false,
  "recommended_action": "[What to do with this feedback]",
  "quotable": "[Best quote for testimonial, or null if negative/neutral]"
}

Scoring Guide:
- sentiment_score: -1.0 (very negative) to 1.0 (very positive), 0 is neutral
- Extract emotions like: satisfied, frustrated, confused, delighted, disappointed, angry, hopeful, indifferent

churn_risk indicators:
- high: Mentions leaving, canceling, or switching to competitor
- medium: Expresses significant frustration or unmet expectations
- low: Generally positive or constructive criticism

response_recommended should be true if:
- Direct question asked
- Significant complaint that needs acknowledgment
- High-value feedback worth thanking
- Churn risk is medium or high