Feedback

AI Squared allows you to capture direct feedback from business users who interact with AI model outputs embedded through Data Apps. This feedback is essential for evaluating model relevance, accuracy, and user confidence—fueling continuous improvement.

---

​

Types of Feedback Supported

​

1. Thumbs Up / Thumbs Down

A binary feedback option to help users indicate whether the insight was useful.

• ✅ Thumbs Up — Insight was helpful
• ❌ Thumbs Down — Insight was not helpful

---

​

2. Rating (1–5 Scale)

Provides a more granular option for rating insight usefulness.

• Configure number of stars (3 to 5)
• Users select one rating per insight interaction

---

​

3. Text-Based Feedback

Capture open-ended qualitative feedback from users.

• Use for additional context when feedback is negative
• Example: “Prediction didn’t match actual customer churn status.”

---

​

4. Multiple Choice

Provide users with a predefined set of reasons for their rating.

• Example for thumbs down:
  • ❌ Not relevant
  • ❌ Incomplete data
  • ❌ Low confidence prediction

---

​

How to Enable Feedback

1. Go to your Data App > Edit.
2. Scroll to the Feedback Settings section.
3. Toggle ON any of the following:
   • Thumbs
   • Star Ratings
   • Text Input
   • Multi-Select Options
4. Save the Data App.

Feedback will now appear alongside model outputs when embedded in business apps.

---

​

Viewing Collected Feedback

Navigate to:

Reports > Data Apps Reports → Select a Data App

There, you’ll find:

• Feedback submission counts
• % positive feedback
• Breakdown by feedback type
• Most common comments or reasons selected

---

​

Best Practices

• Keep feedback simple and non-intrusive
• Use feedback data to validate models
• Combine with usage metrics to gauge adoption quality

---

​

Next Steps

• 👉 Monitor Usage: Analyze how your AI models are performing based on user activity and feedback.