> ## Documentation Index > Fetch the complete documentation index at: https://docs.getarbol.com/llms.txt > Use this file to discover all available pages before exploring further. # AI Employee Analytics > Understanding performance metrics and analytics for individual AI employees # AI Employee Analytics ## Table of Contents 1. [Accessing Employee Analytics](#accessing-employee-analytics) 2. [Key Performance Metrics](#key-performance-metrics) 3. [Daily Activity Analysis](#daily-activity-analysis) 4. [Status Distribution Charts](#status-distribution-charts) 5. [Hourly Distribution Patterns](#hourly-distribution-patterns) 6. [Analysis Tab and Conversation Review](#analysis-tab-and-conversation-review) 7. [Filtering and Searching Conversations](#filtering-and-searching-conversations) 8. [Evaluation Tracking](#evaluation-tracking) 9. [Using Analytics for Optimization](#using-analytics-for-optimization) 10. [Reporting and Insights](#reporting-and-insights) *** ## Accessing Employee Analytics Each AI employee has a dedicated analytics view providing detailed insights into their performance. ### Navigation 1. Go to **AI Employees** from the main navigation 2. Click on the AI employee name you want to analyze 3. The default view shows the **Overview** tab with summary analytics 4. 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 **Context matters**: A new AI employee will have lower total calls than an established one. Compare against similar timeframes and use cases. ### 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 **Outbound Calls**: Conversations initiated by the AI employee calling contacts. High outbound volume indicates: * 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 volumes ### Identifying 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) **Spikes and anomalies**: Sudden peaks may correlate with: * 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 **Typical outbound profile**: * Moderate Success (20-40%) * Significant Voicemail and No Answer * Some Busy during business hours **Warning signs**: * 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 hour ### Identifying Peak Hours Most businesses have predictable peak periods: **B2B typical peaks**: * Late morning (10-11 AM) * Early afternoon (1-3 PM) **B2C typical peaks**: * 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 occurred ### Accessing 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 **Search tips**: * 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 **Direction filter**: Focus on inbound or outbound calls: * All directions * Inbound only * Outbound only **Channel filter**: Filter by communication channel if multiple are used. **Date filters**: Narrow results by time period: * Preset ranges (Last 3 days, Last 7 days, etc.) * Custom date range selection **Duration filters**: Find calls within specific length ranges: * Minimum duration * Maximum duration **Phone number search**: Find calls involving specific numbers. ### 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? **Viewing evaluation results**: Individual conversation details include evaluation scores when available. ### 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? **High failed rates**: Technical investigation priorities: * Check phone number configuration * Review network and connectivity * Examine error logs if available **Unexpected patterns**: Sudden changes warrant investigation: * 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: 1. Capture key metrics at consistent intervals 2. Document trends and significant changes 3. Note actions taken and their results 4. 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 *** ## Summary AI employee analytics provide the visibility needed to continuously improve your voice communication operations. By understanding what each metric means, how to interpret charts, and how to filter for specific insights, you can make data-driven decisions that enhance performance. Regular analytics review, combined with action on insights, creates a cycle of continuous improvement that maximizes the value of your AI employees.