AI Employee Analytics
Table of Contents
- Accessing Employee Analytics
- Key Performance Metrics
- Daily Activity Analysis
- Status Distribution Charts
- Hourly Distribution Patterns
- Analysis Tab and Conversation Review
- Filtering and Searching Conversations
- Evaluation Tracking
- Using Analytics for Optimization
- Reporting and Insights
Accessing Employee Analytics
Each AI employee has a dedicated analytics view providing detailed insights into their performance.Navigation
- Go to AI Employees from the main navigation
- Click on the AI employee name you want to analyze
- The default view shows the Overview tab with summary analytics
- Additional depth is available in the Analysis tab
Analytics Overview
The analytics page presents:- Key metric cards with current performance data
- Visual charts showing trends and distributions
- Comparative data to understand patterns
Date Range
By default, analytics show data from a recent period (typically the last 30 days). This provides meaningful trend data while keeping focus on current performance.Key Performance Metrics
The top of the analytics page displays essential metrics in card format.Total Calls
The total number of conversations this AI employee has handled during the analysis period. What this tells you:- Overall activity level
- Workload compared to other employees
- Capacity utilization
Inbound vs Outbound Breakdown
Understanding the direction of calls provides important context: Inbound Calls: Conversations initiated by contacts calling in. High inbound volume indicates:- The AI employee is serving as a primary contact point
- External awareness of the phone number
- Reactive communication patterns
- Active campaign involvement
- Proactive outreach strategies
- Lead generation or follow-up activities
Success Rate
The percentage of calls that achieved a positive outcome. Calculation: (Successful calls + Voicemail left) / Total calls x 100 Interpreting success rates:- 80%+ for inbound is typically strong
- 20-40% for cold outbound is common
- Compare against your own historical benchmarks
Average Duration
The mean length of conversations, measured in seconds or displayed as minutes:seconds. What duration indicates:- Very short calls might suggest quick resolutions or early disconnects
- Very long calls could indicate complex issues or inefficiencies
- Optimal duration depends on your use case
Total Duration
Cumulative time spent on calls, showing total investment in conversations. Use this metric to:- Understand time commitment
- Plan capacity
- Calculate efficiency ratios
Daily Activity Analysis
The daily activity chart visualizes call patterns over time.Reading the Chart
X-axis: Dates within the analysis period Y-axis: Number of calls per day Visual elements: Bars or lines representing daily call volumesIdentifying Patterns
Consistent volume: Similar call counts each day suggest stable operations. Growth trends: Increasing daily volumes may indicate successful marketing, seasonal demand, or expanding use. Decline patterns: Decreasing volumes warrant investigation - are there configuration issues, or has demand shifted? Weekly cycles: Many businesses see patterns by day of week:- Higher volume Monday-Wednesday
- Lower volume Thursday-Friday
- Minimal weekend activity (for B2B)
- Campaign launches
- Marketing activities
- External events affecting caller behavior
Using Daily Data
Capacity planning: Identify your busiest days to ensure adequate coverage. Campaign timing: Launch outbound campaigns on days with historically better answer rates. Performance investigation: Correlate daily performance drops with specific events or changes.Status Distribution Charts
Visual breakdown of call outcomes helps identify where improvements are needed.Understanding Status Categories
Success: Calls that completed with positive outcomes. The AI employee fulfilled the conversation objective. Voicemail: Calls where voicemail was detected. Depending on configuration, a message may have been left. No Answer: Calls that rang but were not answered. No voicemail system was detected. Busy: The contact’s line was engaged when the call was attempted. Failed: Technical issues prevented call completion. This could include network errors, system issues, or audio problems. Canceled: Calls that were terminated before completion, either programmatically or by user intervention.Visual Representation
Status distribution is typically shown as:- Pie charts showing proportional breakdown
- Bar charts comparing category counts
- Color-coded segments for quick recognition
Analyzing Distribution
Healthy inbound profile:- High Success percentage (70%+)
- Minimal Failed calls
- Low Busy rates
- Moderate Success (20-40%)
- Significant Voicemail and No Answer
- Some Busy during business hours
- High Failed percentage suggests technical investigation needed
- Very low Success on inbound indicates configuration issues
- Unexpected status changes from historical patterns
Hourly Distribution Patterns
Understanding when calls occur helps optimize operations.The Hourly Chart
This visualization shows call volume distributed across hours of the day. X-axis: Hours (0-23 or displayed as AM/PM) Y-axis: Number of calls during each hourIdentifying Peak Hours
Most businesses have predictable peak periods: B2B typical peaks:- Late morning (10-11 AM)
- Early afternoon (1-3 PM)
- Lunch hours
- Early evening (5-7 PM)
- Weekend afternoons
Using Hourly Insights
Optimize outbound timing: Schedule campaigns for hours when answer rates are highest. Plan human backup: Ensure transfer destinations are staffed during peak hours. Identify opportunities: Off-peak hours might offer opportunities for scheduled callbacks.Analysis Tab and Conversation Review
The Analysis tab provides detailed access to individual conversations handled by the AI employee.Conversation List
A table displays all conversations with key information: Title: The conversation subject or automatically generated title Status: Outcome of the conversation Duration: Length of the call Date: When the conversation occurredAccessing Conversation Details
Click any conversation row to navigate to the full conversation detail page, where you can:- Read the complete transcript
- Listen to the audio recording
- View conversation properties
- See evaluation results
Conversation Counts
The header shows the total number of conversations matching your current filters, helping you understand the scope of data you are reviewing.Filtering and Searching Conversations
Powerful filtering helps you find specific conversations or analyze subsets of data.Search Functionality
The search bar allows text-based searching across:- Conversation titles
- Contact names
- Phone numbers
- Use partial matches to find related conversations
- Search for specific topics discussed
- Combine with filters for precise results
Filter Options
Status filter: Select one or more status types to include:- Success
- Failed
- Voicemail
- No Answer
- Busy
- Canceled
- All directions
- Inbound only
- Outbound only
- Preset ranges (Last 3 days, Last 7 days, etc.)
- Custom date range selection
- Minimum duration
- Maximum duration
Property Filters
If you have configured custom properties, these appear as additional filter options:- Select property type
- Choose values to filter by
Active Filter Display
When filters are applied, badges appear showing your active selections:- Click badges to remove individual filters
- Use clear options to reset all filters
Evaluation Tracking
If evaluation criteria are configured for this AI employee, the Analysis tab includes tracking capabilities.Evaluation Criteria Display
A sidebar or panel shows:- All evaluation criteria assigned to this AI employee
- Which criteria are actively being tracked
- Links to configure or modify criteria
Understanding Evaluations
Evaluation criteria are AI-powered assessments that automatically analyze conversation quality:- Did the AI employee achieve the call objective?
- Were required disclosures made?
- Was the tone appropriate?
Tracking Properties
Alongside evaluation criteria, custom properties can be tracked and analyzed:- Properties auto-extracted during conversations
- Manual property assignments
- Property-based filtering and reporting
Using Analytics for Optimization
Analytics exist to drive improvement. Here is how to turn data into action.Identifying Issues
Low success rates: Listen to sample conversations to understand why calls are not succeeding:- Is the prompt inadequate for common questions?
- Are there technical audio quality issues?
- Is the target audience appropriate?
- Check phone number configuration
- Review network and connectivity
- Examine error logs if available
- What changed in configuration recently?
- Are there external factors affecting calls?
- Is there a technical issue?
Configuration Refinement
Use analytics to guide configuration changes: Prompt improvements: If common questions are not being handled well, update the system prompt to address them. Timing adjustments: Schedule campaigns for hours with better historical success. Transfer optimization: If transfers are high, consider whether the AI employee needs expanded capabilities.Comparative Analysis
If you have multiple AI employees:- Compare performance across similar roles
- Identify what top performers do differently
- Apply successful patterns to underperformers
Trend Monitoring
Track metrics over time to identify:- Gradual improvements from optimization efforts
- Degradation that needs attention
- Seasonal patterns to anticipate
Reporting and Insights
Exporting Data
While the platform may not offer direct data export in all areas, you can:- Screenshot charts for presentations
- Use the command palette to ask the AI assistant for specific data summaries
- Navigate to individual records for detailed information
Building Reports
For periodic reporting, develop a standard approach:- Capture key metrics at consistent intervals
- Document trends and significant changes
- Note actions taken and their results
- Plan next optimization steps
Sharing Insights
Use analytics to communicate with stakeholders:- Volume metrics show scale of operations
- Success rates demonstrate effectiveness
- Trend data illustrates improvement over time
AI-Assisted Analysis
Throughout the platform, “Ask AI” links provide access to intelligent analysis:- Request explanations of patterns
- Ask for optimization recommendations
- Get help interpreting complex data