Auto-Labeling
The Auto-Labeling tool allows your bot to analyze messages and automatically tag conversations with the appropriate Labels in realtime. The bot evaluates the conversation against the conditions you define and applies the matching label as soon as a condition is met.
This is useful for organizing conversations without manual tagging, for example, applying a "complaint" label as soon as the bot detects the customer is raising an issue about a product.
You can navigate to Bots → Select your bot → Tools → Preset → Auto-Labeling
Configuration
Define Conditions & Labels
Use this section to define the rules that categorize conversations and determine which label(s) should be applied.
Each row in the table pairs one condition with one or more labels, and consists of the following fields:
-
Description — describe when this label should be applied to the conversation. Be specific about the intent or scenario, rather than relying on exact phrases only.
Example:
When a customer raises a complaint regarding the product. -
Label(s) — select one or more labels to apply when the condition above is matched. Use the dropdown to choose from your existing labels.
Use the "+ Add" button to add another condition/label row.
Each row also has its own trash icon to delete a condition/label pair.
Click Save once you've finished defining your conditions and labels to enable the tool.
How It Looks in Conversation
When a conversation matches a condition you've defined, the bot automatically applies the associated label to the conversation.
The chat will display a system message indicating the label that was applied, for example: "Conversation labeled as complaint by BOT - Car Sales"
Agents can immediately see the assigned labels and use them for follow-up actions, routing, or reporting.
Best Practices
- Write conditions that describe customer intent clearly, so the bot can reliably match real conversations to the right label
- Keep each condition focused on a single scenario rather than combining multiple unrelated triggers
- Review your conditions periodically to avoid overlaps where multiple conditions could match the same conversation
