Freshdesk has four automation systems. They overlap, fire at different events, and have different timing guarantees. Picking the wrong one for the job ends in confusing duplicate actions or actions that never fire.
The four systems
- Dispatcher: runs at ticket create
- Observer: runs on ticket update
- Supervisor: runs on a schedule (every hour by default)
- Scenario Automations: run manually by an agent
- Workflow Automator: visual builder, fires on events plus delays
When to use which
- New ticket triage and assignment → Dispatcher
- Status-change reactions, escalations → Observer
- Time-based “if no response in X hours” → Supervisor or Workflow Automator
- One-click bulk actions an agent invokes → Scenario Automations
- Multi-step branching logic with delays → Workflow Automator
Avoid duplicate firing
If a Dispatcher rule and a Workflow Automator both fire on ticket create with overlapping conditions, both run. Audit conditions carefully. Use distinct tag-based triggers to keep them disjoint.
Dispatcher rule "Tier-1-routing":
IF priority = High AND tag NOT contains "tier1-routed"
THEN assign group, add tag "tier1-routed"
Workflow "Tier-1-followup":
TRIGGER tag added "tier1-routed"
ACTION wait 4h, check status, escalate if open
Supervisor vs Workflow Automator for time delays
Supervisor checks on a fixed cadence (hourly). Workflow Automator can wait an exact duration from a trigger event. For SLA-grade timing, prefer Workflow Automator.
Scenario Automations are not for routine work
If an agent runs the same scenario on every ticket, automate it. Scenarios are for the 20% of cases that need human judgment plus a multi-step action.
Debugging fired vs not fired
Each automation has an execution log. When a rule “should have fired” but didn’t, check the log first. Common reasons: condition mismatch on case sensitivity, tag race conditions, rule disabled by an admin during a change window.
GET /api/v2/automation/{type}/{id}/executions?limit=50
What to do this week
Audit your active rules across all four systems, document which fires when, eliminate condition overlap, and review the execution logs for any rule that was “supposed” to fire last week.