AI Agent tagsUpdated 2 hours ago
Who can use this feature?
Owner, Admin and Lead roles
All plans with AI Agent
Can be used with Shopify stores connected to Gorgias
As AI Agent handles tickets, it automatically applies tags to reflect what it did and where the ticket stands. These tags power the AI Agent ticket view in your helpdesk inbox, and you can use them as conditions in rules to route, prioritize, or automate follow-up on AI Agent tickets.
This article covers:
Lifecycle tags
Lifecycle tags are applied by AI Agent as it processes a ticket. They reflect the current state of the ticket and AI Agent's final outcome.
| Tag | What it means | When it's applied |
|---|---|---|
ai_processing | AI Agent is actively working on the ticket | Applied when AI Agent picks up a ticket. Removed once AI Agent reaches a final outcome. |
ai_answered | AI Agent sent at least one reply to the customer | Applied whenever AI Agent responds, regardless of whether the ticket was ultimately resolved or handed over |
ai_snooze | AI Agent is waiting for the customer to respond | Applied when AI Agent sends a reply that requires follow-up — the ticket is snoozed for 3 days. If the customer doesn't reply, AI Agent closes the ticket |
ai_close | AI Agent resolved the ticket | Applied when AI Agent fully handles the conversation and closes the ticket without human involvement |
ai_handover | AI Agent escalated the ticket to your team | Applied when AI Agent hands the ticket over — because it lacked confidence, detected frustration, encountered a handover topic, or the customer asked to speak with a human |
ai_ignore | AI Agent skipped the ticket entirely | Applied when a ticket matches the conditions in your Prevent AI Agent from answering rule. AI Agent takes no action — no reply, no handover |
Note: ai_answered and ai_handover can appear on the same ticket. This happens when AI Agent replies during the conversation but ultimately escalates rather than resolving it.
Support actions tags
Support actions are automations you configure to make changes in other apps — for example, cancelling an order, updating a shipping address, or processing a return. These tags are applied when AI Agent uses or attempts to use one of your enabled actions.
| Tag | What it means | When it's applied |
|---|---|---|
ai_executed_action | AI Agent successfully completed a support action | Applied when AI Agent uses a support action and it succeeds |
ai_failed_action | AI Agent attempted a support action but it failed | Applied when the action couldn't be completed — for example, if the order was already shipped or something went wrong with the connection to the app AI Agent was trying to reach |
You can create a rule that triggers on ai_failed_action to automatically assign those tickets to a specific team or flag them for urgent follow-up. To investigate why an action failed, go to AI Agent > Support Actions, select the action, and click View Events to see the request log. Learn more: View event logs for AI Agent actions.
The AI Agent ticket view
When you set up AI Agent, Gorgias creates a dedicated AI Agent ticket view in your helpdesk inbox. The view is organized into sections that correspond to AI Agent's lifecycle tags:
| View section | Powered by |
|---|---|
| All | All AI Agent tickets |
| Processing | ai_processing |
| Snooze | ai_snooze |
| Close | ai_close |
| Handover | ai_handover |
| Ignore | ai_ignore — only visible when the Prevent AI Agent from answering rule is installed |
| To review | Not tag-based — see below |
To review
The To review section collects tickets where AI Agent sent a response in one place, so you don't have to search for them. From any ticket in this view, you can rate AI Agent's response, give feedback on the knowledge sources it used, and create new guidance when something could be better.
Unlike the other sections, "To review" isn't controlled by a tag — Gorgias populates it automatically based on AI Agent's involvement in the ticket. Learn how to coach AI Agent.
Use tags in rules
Because AI Agent tags are applied automatically, they're reliable conditions for rules that target AI Agent tickets. For example, you can route ai_handover tickets to a specific team, escalate ai_failed_action tickets, or use the Prevent AI Agent from answering rule to apply ai_ignore based on conditions you define.
Learn more: Use rules with AI Agent (link TBD)
