AI Command Center - User Guide

Access the Dashboard

• Navigate to the AI Command Center from your main menu
• You'll see an overview of all your AI features displayed as cards

Understanding the Overview

• Each card represents one AI feature (called a “use case”)
• Cards show key information like usage count, cost, and status
• Green status means the feature is active, gray means inactive

Viewing AI Performance

Check Overall Performance

• Look at the dashboard cards for a quick overview
• Use the date filter (Today, Last 7 days, etc.) to see different time periods
• Sort cards by name, status, usage, or cost using the dropdown menu

Dive Deeper into Specific Features

• Click on any AI feature card
• View detailed charts showing usage trends and performance
• See cost breakdowns and user activity

Monitoring AI Features

View Feature Status

• Check which features are active or inactive
• Monitor feature availability to users
• Track feature performance and usage patterns

Analyze Feature Performance

• Review execution success rates
• Check user feedback scores
• Identify features that may need attention

Monitor and Debug AI Calls

• View complete details of every AI call made
• Inspect request and response data
• Track execution time and token usage
• Debug failed calls with detailed error messages
• Filter calls by status, user, or time period

Monitoring Usage and Costs

Track Spending

• View cost information on each feature card
• Click into detailed view for cost breakdowns
• Monitor trends to identify expensive features

View Activity Logs

• Click on any feature to see its execution history
• Filter logs by date, user, or success/failure
• Use this to troubleshoot issues

Reviewing Chat Conversations (for Chat Agents)

Access Chat History

• Select a Chat Agent type feature
• Click “View Chat Sessions”
• Browse conversation history with customers

Analyze Chat Performance

• Review how well the AI responded
• Check customer feedback scores
• Identify areas for improvement

Organizing Your Features

Reorder Features

• Drag and drop cards to arrange them in your preferred order
• Put most important features at the top

Search and Filter

• Use the search box to quickly find specific features
• Apply filters to show only active or inactive features

About Usage and Performance

Q: How do I know if my AI features are working well?

A: Look for high success rates (shown in green) and positive user feedback scores. If you see many failures or low feedback, investigate the execution logs.

Q: Which AI features are most popular?

A: Sort your dashboard by “Executions” to see which features are used most often. Features with higher usage numbers are more popular.

Q: How can I see who's using AI features?

A: Click into any feature's detailed view to see user activity. You'll find lists of top users and usage patterns.

About Costs

Q: Why are my AI costs high?

A: Check the detailed cost breakdown in each feature. High token usage or frequent executions increase costs. Consider if the feature provides enough value to justify the expense.

Q: What's the difference between LLM costs and fixed costs?

A: LLM costs are based on actual AI usage (tokens processed). Fixed costs are set amounts you assign to features for internal accounting.

Q: Can I set spending limits?

A: Currently, you can monitor costs and deactivate expensive features manually. Set up regular reviews to track spending trends.

About Monitoring

Q: What does it mean when a feature shows as inactive?

A: An inactive status means the feature is currently disabled and not available to users. Contact your administrator if you need to change feature availability.

Q: How long is monitoring data retained?

A: Monitoring data is retained according to your organization's data retention policy. Check with your administrator for specific retention periods.

Q: How often should I check the Command Center?

A: Check weekly for general monitoring, daily if you're actively tracking costs or troubleshooting issues.

Q: I see failed executions - what should I do?

A: Click into the feature's execution logs, filter by “Failed” status, and review error details. Common issues include connectivity problems or configuration errors.

Q: How do I improve AI performance?

A: Monitor user feedback scores and execution success rates. Low scores may indicate the need to adjust prompts or settings in the Prompt Builder.

Q: The dashboard is loading slowly - is this normal?

A: Performance may vary with large amounts of data. Use date filters to view smaller time periods, or contact support if problems persist.