Settingsintermediate

Priority band targets explained

The four priority bands — low, normal, high, urgent — are not agent-assigned priorities. They're time tripwires the timer crosses as it counts down. Understanding this is the difference between using SLAs effectively and arguing with the badge.

4 min read

Priority band targets explained

The single most important thing to understand about SLA policies in Atender: priority bands are not priorities an agent assigns. They are time tripwires the SLA timer crosses as it counts toward breach.

The model

Each metric type in a policy defines four time targets, all measured in seconds (or hours, in the UI) from when the timer starts:

  • Low — low — Earliest point at which the conversation is “still healthy”
  • Normal — normal — When the conversation crosses from green into At risk
  • High — high — When the conversation crosses from At risk into Urgent
  • Urgent — urgent — When the conversation breaches

So for a First Reply Time metric:

  • Low — 0 — Timer starts. Badge: SLA (green)
  • Normal — 1 hour — Crosses to At risk. Badge: yellow
  • High — 4 hours — Crosses to Urgent. Badge: red
  • Urgent — 24 hours — Breach. Badge: red, counting up

A conversation that gets a reply at 30 minutes shows SLA / Met. One that gets a reply at 3 hours shows At risk / Met (note: the worst-state-reached rule means the badge stays At risk even after the agent replies). One with no reply at all hits Urgent at 4 hours and Breach at 24 hours.

Why “priority bands” rather than just thresholds

The four-band model serves two purposes:

  1. Visual escalation — agents see the urgency before breach, not just after.
  2. Automation hooks — the SLA at risk automation trigger fires when a conversation enters At risk or Urgent (i.e., when it crosses Normal or High). This gives you a chance to intervene before breach. See the SLA escalation recipe.

Without intermediate bands, you’d only know about an SLA problem at the moment it became too late.

Picking sensible band values

The most common shape:

  • Low — 0 — 0h (timer starts here)
  • Normal — 1/3 to 1/2 — 8–12h
  • High — 2/3 to 3/4 — 16–18h
  • Urgent — 1.0 — 24h (this is where breach lives)

The split between Low/Normal sets when At-risk warnings start nagging you. The split between Normal/High sets when escalations should fire. The Urgent value is the actual SLA promise.

If the bands are too tight (Normal at 90% of Urgent), you barely get warning before breach. If they’re too loose (Normal at 10%), every conversation looks At-risk all the time and the signal becomes noise.

The bands are per-metric, per-policy

If a policy tracks both First Reply (24h target) and Resolution (5d target), the bands for each are configured independently:

  • First Reply: Low=0, Normal=8h, High=16h, Urgent=24h
  • Resolution: Low=0, Normal=2d, High=4d, Urgent=5d

A conversation can be in Urgent on First Reply (running out of time to first respond) while still in SLA on Resolution (plenty of room before total resolution would breach). The conversation’s overall worst-state badge is Urgent.

What “priority” doesn’t mean here

Worth saying explicitly:

  • These are NOT severity levels you’d see on a ticketing system (P0, P1, P2, P3).
  • An agent does NOT assign a priority on a conversation that picks a band.
  • The band is a function of time elapsed, not of the conversation’s importance or category.

If you want priority-style differentiation (e.g., VIP conversations get a faster SLA than standard ones), you do it by assigning a different policy to those conversations — typically through customer tier linkage or team/channel scoping. Same band model, different durations.

See also

Tags

Reference
Back to Settings