GovEagle includes a built-in feedback system that helps continuously improve AI Assistant responses. Users can rate individual AI messages with a simple thumbs up or thumbs down, and administrators can review this feedback in the Feedback Center to identify patterns, improve AI behavior, and refine organizational rules.
How User Feedback Works
Every response from the GovEagle AI Assistant includes feedback buttons that allow users to quickly rate the quality and helpfulness of that specific message.
Providing Feedback as a User
When the AI Assistant responds to your request in the Chat interface, you'll see thumbs up and thumbs down icons below the message.
Feedback buttons
To provide feedback on an AI response:
Click the thumbs up icon if the response was helpful, accurate, and met your needs
Click the thumbs down icon if the response was unhelpful, inaccurate, or missed the mark
After clicking either button, a feedback dialog appears.
Feedback dialog
The feedback dialog allows you to provide additional context:
Heading: The dialog is titled "Feedback" and notes "This will be shared with admins to improve AI responses"
Feedback field: Enter details explaining why you liked or disliked the response (up to 500 characters)
Submit or Cancel: Click Submit to send your feedback to administrators, or Cancel to close the dialog without submitting
What to Include in Your Feedback
Helpful feedback is specific and actionable. Consider including:
For positive feedback (thumbs up):
What made the response particularly useful
Specific elements that were well done
How the response helped you accomplish your task
For negative feedback (thumbs down):
What was missing or incorrect
How the response failed to meet your needs
What you expected instead
Specific errors or problems you noticed
Examples of helpful feedback:
Good: "Perfect! The outline included all required sections from the RFP and mapped them correctly to evaluation criteria."
Less helpful: "Good job!"
Good: "The AI referenced past performance from a proposal written by our teaming partner. We should only use past performance from proposals where GovEagle was the prime contractor."
Less helpful: "This is wrong."
Why Your Feedback Matters
User feedback helps your organization:
Identify areas where the AI Assistant needs improvement
Recognize when AI responses are particularly effective
Inform decisions about Organization Rules and Writing Style Guides
Provide data to GovEagle support when requesting AI updates
Continuously improve the quality of AI assistance for everyone
The Feedback Center (For Administrators)
Administrators can access the Feedback Center to review all feedback submitted by users across the organization.
Accessing the Feedback Center
To view user feedback:
Navigate to Settings from the left sidebar
Select Feedback from the settings menu
Feedback Portal
Understanding the Feedback Table
The Feedback Center displays all submitted feedback in a sortable table with the following columns:
Date
When the feedback was submitted. Click the down arrow to sort by date (newest to oldest or vice versa).
User
The name of the user who provided the feedback. This helps you identify patterns among specific users or roles.
User Rating
Indicates whether the user gave positive feedback (thumbs up) or negative feedback (thumbs down). Positive ratings appear in green with an upward arrow (△ Positive), while negative ratings appear in red with a downward arrow (▽ Negative).
User Feedback
The detailed feedback text the user provided explaining their rating. Some entries may show "No feedback text" if the user submitted a rating without additional comments.
User Prompt
The question or request the user sent to the AI Assistant that generated the rated response. This provides context for understanding the feedback.
Agent Message
The beginning of the AI Assistant's response that the user rated. This column is truncated but shows enough to identify which response received the feedback.
Analyzing Feedback Patterns
When reviewing feedback, look for:
Common complaints or issues:
Do multiple users report similar problems with certain types of requests?
Are there consistent gaps in AI responses?
Do users frequently need to clarify or rephrase requests?
Positive patterns:
What types of responses do users consistently rate highly?
Which features or capabilities are working well?
Are there best practices you can identify and share?
User-specific patterns:
Does one user have many negative ratings? They may need additional training or support.
Are certain users particularly satisfied? They might be good candidates to share best practices.
Topic-specific patterns:
Do certain proposal tasks generate more negative feedback (e.g., outlines, drafting)?
Are there specific domains where the AI needs more context?
Taking Action on Feedback
Based on the feedback you review, you can take several actions:
1. Create or Update Organization Rules
If feedback reveals consistent issues or preferences, you can address them through Organization Rules in Settings > AI Preferences.
Example: If users report that the AI sometimes references past performance from teaming partner proposals, you might add an Organization Rule:
When referencing past performance, only use proposals where GovEagle served
as the prime contractor. Do not reference proposals written by our teaming
partners or where we served as a subcontractor unless explicitly requested.
2. Update Your Writing Style Guide
If feedback indicates the AI's writing style doesn't match expectations, update your Writing Style Guide in Settings > AI Preferences.
Example: If users consistently ask for more concise responses, you might add:
Keep responses concise and direct. Start with the most important information
first and provide additional detail only when relevant.
3. Provide User Training
If feedback suggests users aren't getting the most from the AI, consider:
Sharing best practices for writing effective prompts
Demonstrating successful workflows in team meetings
Creating internal documentation with examples
Highlighting particularly effective user prompts as examples
4. Request Product Updates
For issues that can't be resolved through rules or training, you can contact GovEagle support. The feedback data provides concrete examples to share with the product team.
When contacting support, include:
Specific feedback entries that illustrate the issue
How frequently the problem occurs
The business impact of the issue
Any Organization Rules or settings you've already tried
5. Follow Up with Users
For particularly helpful or concerning feedback, consider reaching out to the user directly:
Thank users who provide detailed, constructive feedback
Clarify feedback that's unclear
Help users who are struggling to get good results
Learn more about specific use cases
Best Practices for Reviewing Feedback
Review Regularly
Set a schedule for reviewing feedback (weekly, biweekly, or monthly depending on usage volume). Regular reviews help you identify trends before they become significant problems.
Look for Trends, Not Outliers
A single negative review may not indicate a systemic problem. Focus on patterns that appear across multiple users or over time.
Balance Positive and Negative
Don't only focus on negative feedback. Positive feedback shows what's working well and should be reinforced or replicated.
Share Insights
Communicate feedback trends with your team. Let users know when their feedback leads to improvements—this encourages continued participation.
Track Improvements
After making changes based on feedback (new rules, updated style guides), monitor whether subsequent feedback improves. This helps you measure the effectiveness of your interventions.
Respect User Privacy
While feedback is shared with administrators, treat it respectfully. Use it to improve systems and processes, not to criticize individual users.
Encouraging Meaningful Feedback
As an administrator, you can encourage users to provide helpful feedback:
Communicate the Value
Let users know that their feedback directly improves the AI for everyone. Share examples of improvements made based on user feedback.
Make It Easy
The feedback system is designed to be quick and simple. Remind users they can submit ratings with or without detailed comments—both are valuable.
Respond to Feedback
When users see their feedback leading to tangible improvements, they're more likely to continue providing it. Communicate changes you've made based on user input.
Set Expectations
Help users understand what kinds of feedback are most useful:
Specific rather than general
Descriptive rather than evaluative
Focused on the AI's response rather than external factors
Feedback and Continuous Improvement
The feedback system is part of GovEagle's commitment to continuous improvement. The cycle works like this:
Users work with the AI and rate responses
Feedback is collected in the Feedback Center
Administrators review feedback and identify patterns
Changes are implemented through rules, style guides, or product updates
Users experience improved AI responses
The cycle repeats with ongoing feedback
This continuous improvement cycle ensures the GovEagle AI Assistant becomes increasingly effective for your organization's specific needs over time.
Privacy and Data Use
User feedback is:
Visible to administrators in your organization through the Feedback Center
Used to improve AI responses for your organization
Shared with GovEagle when you request product support or updates
Never shared publicly or with other organizations
Users should feel comfortable providing honest, constructive feedback knowing it will be used solely for improvement purposes.
More questions?
If you have questions about interpreting feedback or taking action on feedback trends, contact GovEagle support for guidance.