> ## Documentation Index
> Fetch the complete documentation index at: https://docs.keloa.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# SLA timers

> Set response and resolution targets per priority, surface conversations approaching breach, and keep your promises.

A **service-level agreement** (SLA) defines how fast you promise to reply and resolve. Keloa's SLA layer attaches a target to every conversation based on its priority, runs the clock, and pulls anything close to breach into a dedicated inbox view so it gets the attention it needs.

<Note>
  SLA policies are available on the **Business** plan and above (paywall key `tickets.use_sla`). Lower plans see the priority field but no timers attach.
</Note>

## How the timer works

Two clocks run on every conversation:

* **First response** — starts when the customer's first message lands, stops when any team member (or the AI) sends a reply.
* **Resolution** — starts at the same point, stops when the conversation is resolved.

When a clock crosses 75% of its target, the conversation surfaces in the inbox **SLA** filter. At 100% it's flagged as **breached** and stays flagged for analytics.

## Configure your policy

Open **Settings → SLA policy**. You'll see a 4-by-2 grid: one row per priority (`urgent`, `high`, `normal`, `low`), one column for first-response, one for resolution. Values are in minutes.

Defaults match what most SMB support teams promise:

| Priority   | First response | Resolution |
| ---------- | -------------- | ---------- |
| **Urgent** | 15 min         | 2 hours    |
| **High**   | 1 hour         | 8 hours    |
| **Normal** | 4 hours        | 24 hours   |
| **Low**    | 24 hours       | 72 hours   |

Two policy switches:

* **Respect business hours** — pause the clock outside your configured business hours so an after-hours email doesn't breach overnight. See [Workspace settings](/settings/workspace) for hours setup.
* **Pause on pending** — pause the clock when a conversation is waiting on the customer (tagged `pending` or `waiting-on-customer`). Resumes when they reply.

## Setting priority on a conversation

Open any conversation and click the priority pill in the header to change it. Priority can also be set automatically:

* By an [assignment rule](/settings/assignment-rules) (e.g. *"VIPs get high priority"*).
* By a flow's **Set priority** action.
* By the AI agent if the system prompt says so.

If no priority is set, conversations default to **normal**.

## The SLA inbox view

A built-in **SLA** filter in the left rail shows everything within 25% of breach, ordered by *time-to-breach* — the most urgent ticket is on top, every refresh.

Your team works the SLA queue first thing in the morning and last thing before signing off. Easy ritual, no missed promises.

## Breaches

A breach doesn't pause anything. The conversation stays in the queue until it's replied/resolved, just with a red **SLA breached** badge.

Breach data shows up in [Analytics → Response time](/analytics/response-time) so you can see whether the policy is realistic.

## Plan availability

| Plan     | SLA timers                          |
| -------- | ----------------------------------- |
| Starter  | Off                                 |
| Growth   | Off                                 |
| Business | On (default policy + customisation) |
| Scale    | On                                  |

Upgrade in **Settings → Billing** if your team needs SLAs.

## Related

<CardGroup cols={2}>
  <Card title="Custom inbox views" icon="filter" href="/inbox/custom-views">
    Build queues like "VIP unassigned over 1h".
  </Card>

  <Card title="Response time analytics" icon="clock" href="/analytics/response-time">
    Where the time goes — by channel, agent, hour of day.
  </Card>
</CardGroup>