> ## Documentation Index
> Fetch the complete documentation index at: https://docs.chatros.com/llms.txt
> Use this file to discover all available pages before exploring further.
# Analytics
> Track performance metrics and monitor your AI Agent activity
The Analytics section provides insights into your AI Agent's performance, conversation volumes, response times, and customer satisfaction. Use these metrics to optimize your agent and improve customer experience.
## Overview Dashboard
### Summary Cards
Four key metrics are displayed at the top of the Analytics page:
**Total Messages**
* Total number of messages sent and received in the selected time period
* Includes both customer messages and AI responses
* Shows aggregate activity across all conversations
**Total Conversations**
* Number of unique conversations started in the time period
* Each conversation represents a customer interaction session
* Tracks how many customers engaged with your agent
**Average Response Time**
* Mean time taken by your AI Agent to respond to customer queries
* Measured in seconds or milliseconds
* Lower response times indicate better performance
**Customer Satisfaction**
* Aggregate satisfaction score based on customer feedback
* May include ratings, thumbs up/down, or other feedback mechanisms
* Helps identify areas for improvement
## Time Range Filter
Select different time periods to view analytics:
| Time Range | Description |
| ---------------- | ------------------------------- |
| **Last 7 Days** | Recent activity and trends |
| **Last 30 Days** | Monthly performance overview |
| **Last 90 Days** | Quarterly insights and patterns |
When you change the time range:
* All metrics update automatically
* Charts redraw with new data
* Summary cards recalculate
## Message Volume Over Time
### Line Chart Visualization
This chart displays message activity trends:
* **X-axis**: Dates within selected time period
* **Y-axis**: Message count
* **Trend Line**: Shows message volume over time
### What to Look For
**Increasing Trend**
* Growing user engagement
* Successful deployment
* Expanding user base
**Decreasing Trend**
* Reduced engagement (investigate causes)
* Seasonal patterns
* Technical issues
**Spikes**
* Marketing campaigns
* Product launches
* Support incidents
**Valleys**
* Weekends or holidays
* Downtime
* Seasonal lulls
## Daily Activity Distribution
### Bar Chart Visualization
Shows conversation distribution by day of the week:
* **X-axis**: Days (Monday - Sunday)
* **Y-axis**: Activity count
* **Bars**: Height represents volume
### Insights
Use this chart to:
* Identify peak activity days
* Plan support staffing
* Schedule agent maintenance during low-traffic periods
* Understand user behavior patterns
**Example Patterns:**
* B2B agents often see weekday peaks
* Consumer agents may be busier on weekends
* Support queries spike on Mondays
## Using Analytics Data
### Performance Monitoring
**Response Time**
* Target: Under 2 seconds for most queries
* If slow: Check training data size, API latency, or server resources
* Optimize by reducing complex embeddings or using faster models
**Conversation Volume**
* Track growth over time
* Compare to website traffic or marketing campaigns
* Identify successful content or features
**Message Volume**
* High messages per conversation = engaged users or complex queries
* Low messages per conversation = quick resolutions or user drop-off
* Balance depends on your use case
### Optimization Strategies
**Improving Satisfaction**
1. Review negative feedback in Activity section
2. Revise AI responses that received poor ratings
3. Save improved answers as Q\&A pairs
4. Monitor satisfaction score improvement
**Reducing Response Time**
1. Optimize training data (remove duplicate or irrelevant content)
2. Use more efficient AI models
3. Implement caching for common queries
4. Review server performance
**Increasing Engagement**
1. Improve agent visibility on your website
2. Use suggested messages to prompt user interaction
3. Deploy agent on high-traffic pages
4. Optimize initial messages for clarity
## Comparing Site Connections
If you have multiple site connections:
* Filter analytics by specific site
* Compare performance across sites
* Identify which sites have the most active conversations
## Data Refresh
* Analytics data updates in real-time for recent activity
* Historical data is aggregated for performance
* Time range changes fetch new data immediately
Analytics tracks all conversations, including both AI and manual modes. Use this data to understand total customer support volume.
## Best Practices
**Daily Monitoring**
* Check Total Messages and Conversations daily
* Look for anomalies or unexpected drops
* Respond quickly to performance issues
**Weekly Reviews**
* Review average response time trends
* Identify days with highest activity
* Plan capacity accordingly
**Monthly Analysis**
* Compare satisfaction scores month-over-month
* Analyze message volume growth
* Identify long-term trends
**Using Data for Training**
* High message volume with low satisfaction = training gaps
* Conversations with many back-and-forth messages = unclear responses
* Review Activity logs for specific examples to improve
## Exporting Data
While the dashboard provides visual insights, you can:
* Take screenshots of charts for reports
* Access detailed conversation logs in Activity section
* Use data to justify support investments or agent improvements
Analytics data is specific to each agent. If you have multiple agents, view analytics separately for each to understand individual performance.
***
## What's Next?
View detailed conversation logs
Improve training data based on insights
Adjust agent configuration
Learn more about optimizing your agent