Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.kontext21.com/llms.txt

Use this file to discover all available pages before exploring further.

​
Message Feedback

Your feedback helps improve X21’s responses and capabilities. Rate AI messages and provide comments to influence future behavior.

​
Rating Responses

​
Thumbs Up / Down

Every AI response includes feedback buttons: Thumbs Up 👍: Response was helpful
  • Accurate results
  • Clear explanations
  • Solved your problem
Thumbs Down 👎: Response needs improvement
  • Incorrect results
  • Unclear explanations
  • Didn’t address request

​
When to Rate

Rate responses when:
  • Result is particularly good or bad
  • You want to reinforce correct behavior
  • Something unexpected happens
  • You have specific feedback

​
Providing Comments

​
Adding Feedback Text

After clicking thumbs down:
  1. Comment box appears
  2. Describe the issue
  3. Be specific about what went wrong
  4. Click submit

​
Effective Feedback

Be specific: 
✓ "The formula calculated incorrectly - used SUM instead of AVERAGE"
✗ "Wrong answer"
Describe expectations: 
✓ "Expected bold formatting on headers, got italic instead"
✗ "Formatting was wrong"
Note what worked: 
✓ "The data analysis was correct, but the chart recommendation didn't match our style guide"

​
What Happens to Feedback

​
Analytics Integration

Feedback is sent to Langfuse analytics:
  • Score: +1 (thumbs up), -1 (thumbs down)
  • Comment: Your text feedback
  • Context: Conversation metadata
  • Timing: When feedback was provided

​
Improvement Cycle

Your feedback helps:
  1. Identify patterns in successful responses
  2. Detect issues with specific operations
  3. Train improvements to the system
  4. Prioritize features based on user needs

​
Feedback Best Practices

​
When to Provide Detailed Feedback

Detailed comments are most valuable for:
  • Unexpected behavior
  • Incorrect results
  • Confusing explanations
  • Missing features

​
Privacy

Feedback includes:
  • Your conversation context
  • Tool operations performed
  • Error messages (if any)
  • Your email (for follow-up)
Don’t include:
  • Sensitive data values
  • Personal information
  • Confidential details

​
Following Up

For critical issues:
  1. Provide detailed feedback
  2. Contact support with conversation ID
  3. Include steps to reproduce

​
Feedback Types

​
Positive Feedback

When to give thumbs up:
  • Perfect first-try results
  • Creative problem solving
  • Clear explanations
  • Time saved

​
Constructive Feedback

When to give thumbs down with comments:
  • Minor issues to fix
  • Suggested improvements
  • Alternative approaches
  • Feature requests

​
Bug Reports

For serious issues:
  1. Thumbs down
  2. Detailed comment
  3. Contact support
  4. Include conversation link

​
Related Topics