AI analytics
See how many messages the AI answers, how often it hands over to a person and how many tokens it uses.
AI Analytics shows how the assistant is performing on the selected WhatsApp number.
Metrics
| Metric | Meaning |
|---|---|
| Total Message | Customer messages received |
| AI Responses | Messages the AI answered |
| Human Escalations | Chats the AI handed to a person |
| Tokens Used | OpenAI tokens used by the assistant |
Switch between Weekly, Last 30 Days and Yearly to change the period. The chart shows the same figures over time.
Reading the numbers
- A rising escalation rate usually means customers are asking about things the AI doesn’t know yet. Look at those chats and add the missing answers in AI Training.
- Tokens Used drives your OpenAI cost and counts towards any AI token limit in your package. Lowering Max Tokens in AI Settings keeps replies shorter.
Clearing AI memory
The AI remembers recent messages in each conversation for context. You can clear that memory for one customer, or for all conversations at once — useful after a big change to your training data or prompt. Customers’ messages and chat history in the inbox aren’t affected.
Still have questions?
Contact usOur team is happy to walk you through Wazline and help you set up your workspace.