AI Feedback Loop Design

Format: Design how users can provide feedback on AI output.

Feedback Mechanisms:

Simplest: 👍/👎 buttons
Intermediate: Star rating (1-5)
Advanced: Multi-dimensional rating (accuracy, usefulness, completeness)

After collecting feedback:
- Regularly review low-score feedback
- Use high-score examples to improve prompts
- Identify systemic issues (certain types of questions consistently get low scores)

Exercise: Design a feedback UI for your AI feature, ensuring:

• Feedback takes no more than 2 clicks
• It doesn't interrupt the user's primary flow
• There's a way to view feedback data trends