Summary
Astra provides detailed analytics for each AI agent, helping you understand how well an individual agent is performing across your website and messaging channels. Use these reports to track conversation outcomes, measure business impact, identify knowledge gaps, and find opportunities to improve your agent's performance.
Instructions
The Agent Analytics page gives you a detailed view of how a specific AI agent is performing. Unlike the main Astra Dashboard, which shows account-wide metrics, Agent Analytics focuses on a single agent.
You can use these reports to:
Monitor conversation performance
Measure how effectively the agent resolves customer inquiries
Understand business impact and time savings
Identify common customer questions
Discover knowledge gaps and escalation trends
Access agent analytics
To view analytics for a specific AI agent:
Log in to your Astra account.
Navigate to Agents from the left-hand menu.
Select the agent you want to analyze.
Open the Analytics section.
The analytics page will display performance data for the selected agent.
Review Astra's business impact
The Astra Impact section provides a high-level overview of how the selected agent contributes to your business.
1. Conversations handled
Shows the total number of conversations handled by the agent across all supported channels during the selected timeframe.
2. Time saved
Estimates the amount of time saved by automating conversations that would otherwise require assistance from a human support team member.
3. Resolution rate
Shows the percentage of conversations successfully resolved without human intervention.
4. Languages handled
Displays the number of unique languages in which the agent successfully supported users.
Analyze conversation performance
The Overview Report and Conversation Funnel provide a detailed breakdown of how conversations are handled from start to finish.
1. Core metrics
These metrics help you evaluate the overall effectiveness of the agent.
Total conversations
The total number of conversations handled during the selected period.
Resolution rate
The percentage of conversations resolved without escalating to a human team member.
Average handle time
The average time required to resolve a conversation, measured in minutes.
Drop-off rate
The percentage of users who leave the conversation before receiving a helpful answer or after expressing dissatisfaction with the AI's response.
2. Understand conversation trends and outcomes
The analytics page also includes reports that help you understand customer behavior and identify improvement opportunities.
Conversation funnel
The conversation funnel shows how conversations progress through different stages:
Total conversations
Resolved conversations
Escalated conversations
Dropped conversations
This helps you understand where conversations are succeeding and where users may be leaving the experience.
Conversation volume trend
Displays conversation activity over time, allowing you to identify trends and peak traffic periods.
You can view performance across:
Text Agent performance
Voice Agent performance
This helps you understand when customers are most actively engaging with your AI agent.
Top FAQ categories
Highlights the topics customers ask about most frequently.
Reviewing these categories can help you:
Understand customer needs
Identify popular support topics
Improve content and training materials
Identify areas for improvement
The analytics page also provides insights into where your AI agent may need additional training or optimization.
1. Knowledge gaps
The Knowledge Gaps report identifies questions or topics the agent could not answer successfully.
Use this information to:
Expand your knowledge base
Add missing information
Improve future response quality
2. Escalation reasons
The Escalation Reasons report shows why conversations were transferred to a human team member.
Common reasons may include:
Complex customer requests
Missing information
Customer dissatisfaction
Workflow or integration limitations
Reviewing escalation patterns can help you improve your agent's training, workflows, and overall customer experience.