Logs & Analytics
Review every conversation, track tool usage, monitor costs, and use data to improve your bot.
Conversation Logs
Access AxiaChat AI → Logs to see all chatbot interactions.
Information Available
- Date and time of each message
- Which bot handled the conversation
- User's message and bot's response
- Session ID
- WordPress user (if logged in)
- Tokens used
- Estimated cost
Filters
- By specific bot
- By date range
- By user or session
- By keyword search
Conversation Detail View
Click on any conversation to see:
- Full session history — all messages exchanged
- Context used for each response
- Token and cost metrics per message
Tool Logs
At AxiaChat AI → Logs → Tools you'll find:
- AI tool invocations (lead capture, appointments, email alerts)
- Webhook executions
- MCP server queries
- Execution duration per call
- Response data received
- Errors (if any)
Usage Metrics
Quick View
- Recent conversations panel
- Daily volume chart
- Most active bots
Global Statistics
- Total messages
- Messages per day / week / month
- Distribution by bot
- Peak activity hours
Cost Tracking
The plugin automatically calculates:
- Input and output tokens per message
- Cost per conversation (based on the model used)
- Accumulated totals
- Time-based comparisons
Export
Export data for external analysis:
- CSV format
- Date range selection
- Filters applied before export
Enable / Disable Logging
- Go to AxiaChat AI → Settings.
- Find the logging option.
- Toggle on or off.
When disabled, no conversations are stored — the chat still works normally but with maximum user privacy.
Using Logs to Improve Your Bot
Identify knowledge gaps
Search for questions the bot couldn't answer well → add that content to your context.
Detect patterns
Find frequently asked questions → add them to your bot's instructions or FAQ context.
Optimize instructions
If responses don't match the desired tone → adjust the system prompt.
Evaluate models
Compare response quality across different models → choose the best one for your use case.
💡 Review weekly
Spend 10 minutes each week reviewing recent conversations. You'll quickly spot patterns that help you improve the bot's training and reduce unhelpful responses.