Labels keep conversations tidy and cut through the noise. You can add a label by hand, or let AI apply it to new conversations. A label that applies itself is an auto-label, and it can also trigger actions like assigning a team or resolving notifications. Auto-labels are available on paid plans.
Labels and auto-labels live on the same page: Settings → Labels. The tabs All, Manual and Auto filter the list.
Creating and editing labels needs the Manager or Administrator role. Agents don't see Settings → Labels. They can still add and remove existing labels on a conversation: in the conversation sidebar, under Labels, click Add.
Add a label
- Go to Settings → Labels and click Add label
- Enter a label name. Keep it short and easy to understand
- Optional: add a Description and pick a Color
- Click Save
Make a label apply itself
- Go to Settings → Labels and add a label, or open an existing one
- Switch on Apply automatically
- In Description, describe the type of message you want to catch, e.g. Customers ask for a refund or mention returning a product
- Choose what happens when the label is applied (see below)
- Click Save
AI reads the subject and body of the first message of each new conversation and applies the one label that fits best. If nothing fits, no label is applied. Later messages in the same conversation aren't checked. It doesn't see the sender's address. To match on the sender, use an automation rule. Conversations from blocked senders (Spam) are not labelled.
What you can set on an auto-label:
-
Show in Inbox: keep these conversations in your main inbox. Untick it to move them to the "Other" view
-
Unassign AI agent: stops the AI agent from replying to conversations with this label
-
Resolve conversation: automatically resolve conversations that don't need a reply, like notifications. Tick Disable CSAT survey if these shouldn't get a survey
-
Assign a team or Assign an agent: send a set of tickets to specific teams or agents
-
Mute conversation or Snooze conversation: for less important messages. A snoozed conversation moves to the Pending status
-
Change priority: help agents spot important messages
-
Add private note: leave an automatic note for your team
Where conversations go when Show in Inbox is unticked
Conversations whose only labels have Show in Inbox unticked are left out of All conversations and your inbox lists. Saved views leave them out too. They keep their status, so they stay open until someone or something resolves them.
To see them, open the Other view under Views in the left sidebar. Commslayer adds this view by itself the first time such a label is applied to a conversation, and everyone on the team can see it. You can also click the label under Labels in the left sidebar.
They can't be deleted. Resolve them from the Other view, or tick Resolve conversation on the label so new ones are resolved as they arrive.
A conversation that also has a manual label, or a label with Show in Inbox ticked, stays in All conversations and your inbox lists.
Auto-labels and your plan
Auto-labels use AI, and your plan includes its own allowance for them, separate from the AI agent and AI drafts. You can follow it in Settings → Usage & billing, in the AI section. Auto-labels stop when the included allowance is used up. They are never billed as pay-as-you-go.
One labelled conversation uses one auto-label from your allowance, however many auto-labels you have. A conversation that gets no label uses nothing.
To use fewer, switch off Apply automatically on the labels you don't need. To stop all of them, switch off Enable auto-labels in Settings → Labels.
On the Free plan auto-labels don't run. Auto-labels, including the built-in ones, still show in the list, but they can't be edited or deleted and they aren't applied to your conversations. Labels without Apply automatically can be edited and deleted on every plan.
Labels and contact reasons
Labels are general-purpose tags that can also trigger automations. Contact reasons are a single classification field per conversation, with their own hierarchy of categories and sub-reasons and their own report. We recommend using both.