Workspace admin

SLA policies and business hours

Updated

Service-level agreements make response and resolution times measurable. IntelliDesk starts an SLA on each ticket from the policy for its priority, then tracks it through the ticket's life: replies, waiting on the customer, closing and reopening.

Two targets per ticket

Every tracked ticket has two due times:

  • First response due: when the first public agent reply must be sent.
  • Resolution due: when the ticket must be closed.

Each is decided once, as met or breached. If no active policy exists for a ticket's priority, the ticket is simply not tracked.

Which tickets get an SLA

Every channel is covered: email, the customer portal, the widget and REST API, WhatsApp, feedback, and tickets agents create themselves.

  • Live chat starts its SLA when the visitor is handed to a person (they ask for one, or the AI hands over). While the AI is handling the chat, nobody on your team is on the clock.
  • Existing tickets are not backfilled. A policy applies to tickets created after it exists, so historic tickets never show up as retroactive breaches.

Creating and editing a policy

Go to Settings → Routing & rules → SLA Policies. Each policy has:

  • Priority: low, normal, high or urgent. Create one policy per priority you want to track.
  • First response and Resolution (in minutes): counted from when the ticket arrives. An 8-hour target is 480.
  • Business hours only: when on, only time inside your business hours counts, using the schedule you pick or your default schedule. When off, the clock runs 24/7.
  • Warn when and Escalate to managers: see Warnings and escalation below.

Use Edit on a policy to change its targets or hours, switch it off (Active), or delete it. Changes apply to new tickets; tickets already in progress keep the targets they started with.

Business hours

Set these up under Settings → Routing & rules → Business Hours: a timezone, the working hours for each day, and holidays.

With a business-hours policy, a ticket that arrives on Friday at 16:30 with a 60-minute first-response target is due at 08:30 on Monday, assuming Mon–Fri 08:00–17:00. Weekends, nights and holidays don't count.

What pauses the clock

The clock pauses while the ticket is waiting on someone else:

  • Pending: you replied and are waiting on the customer. Replying with status Pending pauses the clock.
  • On hold: waiting on a third party.

While paused, the ticket shows Paused instead of a countdown. When the customer replies (or you set the ticket back to Open), the clock resumes with the time it had left, so waiting never eats into your target.

Someone being away, on leave or absent never pauses or moves the clock. Covering for them is the team's job (see Away, on leave or absent: your availability and what happens to your tickets).

Changes along the way

  • Priority change: the ticket switches to the new priority's policy. Its targets are recalculated from when the clock started, and any time it spent paused is still credited. A first response already sent stays met.
  • Closing decides the resolution target: met if closed in time, breached if not.
  • Reopening starts a fresh resolution clock from the moment the ticket reopens. An earlier first-response result stays in the history and in reports, but no longer marks the reopened ticket as breached.

Who counts as the owner

Warnings and escalations go to the ticket's effective owner:

  1. the assignee, if they're available;
  2. otherwise the desk's backup (set in the desk's When someone is on leave or absent card), if they're available;
  3. otherwise the desk's available managers;
  4. otherwise all the desk's managers.

Warnings and escalation

Open a policy with Edit to set both.

Warn when sends a heads-up before a target is missed:

  • Half the time has gone
  • 75% of the time has gone (the default)
  • 90% of the time has gone
  • Don't warn

The warning goes to the effective owner as an SLA warning notification, for example "First response due on HD-1332 in 6 min". It's sent once per due time, measured on the policy's own clock (business hours if the policy uses them). A priority change or a reopen sets new due times, so the warning can fire again.

Escalate to managers raises a ticket that stays open well past its target:

  • Don't escalate (the default)
  • When 25% of the target overdue
  • When 50% of the target overdue
  • When a full target overdue

The escalation goes to all the desk's managers plus the effective owner: "HD-1332 is well past its SLA and still open".

Breaches

A background check runs every 10 minutes. As soon as a due time passes, the target is marked breached and a notification is sent:

  • to the assignee, if they're available;
  • if the assignee is unavailable or the ticket is unassigned, to the desk's backup or available managers and to the desk's managers. The notification notes "(Name is unavailable)" when the owner is out. These breaches are recorded as coverage breaches: they count against the desk, not the person.

A missed first response marks the ticket breached straight away, and resolution keeps being tracked.

Because the check runs every 10 minutes, a warning, escalation or breach notification can arrive up to about 10 minutes after its threshold.

The SLA breached count on the dashboard and the inbox's SLA: Breached filter both show open tickets that missed a target, which are the ones you can still act on. A ticket drops out of the count once it's closed.

Reporting

Reports → SLA shows:

  • First response compliance: every ticket whose first response has been decided, open ones included.
  • Resolution compliance: tickets whose SLA has finished (closed).
  • By priority, by desk and weekly trend: finished SLAs, counted as met only if neither target was missed.
  • Breached tickets: every ticket in the period that missed a target, including open ones.

Tickets in the trash are left out of all reports.

Tips for sane SLAs

  • Start lax, then tighten. Set targets you can comfortably meet, then tighten them as your workflow matures.
  • Match priority to impact, not effort. Urgent means high business impact if delayed, not hard to fix.
  • Don't promise weekend cover you can't deliver. Use business hours only unless someone is actually on call.
  • Use Pending honestly. It pauses your clock, so use it when you genuinely need something from the customer.