How agent tiers work
This example uses one board with two statuses, Tier 1 Queue and Tier 2 Queue.- A new ticket lands in Tier 1 Queue, and the Tier 1 Flow runs its skill.
- The skill assigns a Tier 1 tech, or moves the ticket to Tier 2 Queue if none qualifies.
- The status change is written to the PSA. When the PSA sends that update back, Thread checks the ticket against your Flows again, and the Tier 2 Flow matches and assigns a Tier 2 tech.
Variations
Before you start
- Each tier has its own status (or board), and the statuses are mapped in Thread.
- You’ve listed each tier’s techs.
- The skills are shared skills.
Set up agent tiers
1
Write one skill per tier
Give each skill its pool and rule. In the Tier 1 skill, say exactly when to move a ticket to the Tier 2 status instead of assigning.
2
Create one Flow per tier
Each Flow’s conditions name that tier’s status. Add Agent Builder → Run a Skill with the update ticket and add note tools checked.
3
Test both paths
Send one ticket a Tier 1 tech should take and one that should move up, and check both end with the right owner.
Rules for agent tiers
- One run per ticket per Flow. Re-entering a tier’s status doesn’t re-run that tier.
- Remove the owner to hand a ticket back. Auto-assign only takes unowned tickets, and an agent overwrites an owner unless told not to.
- Statuses still matter for profiles. If any tier uses an Auto-assign profile, the profile rules apply: no shared dispatch statuses on a board, one profile per team, and no fallback loops between profiles.