# SLA
| Item | Details |
|—|—|
| Who it is for | Administrators configure; agents see countdowns / breaches on conversations |
| Menu location | **Settings → SLA**; Reports → SLA |
| Plan dependency | Requires SLA |
SLA (service level agreement) defines how soon you must **first reply**, **next reply**, and **resolve**. Breaches mark the conversation, feed reports, and can notify My Inbox.
—
## 1. Creating a policy
1. **Settings → SLA → Add**.
2. Name it, for example “Standard business hours” or “VIP”.
3. Set each threshold (minutes / hours). Whether only business hours or calendar time count depends on the form — align with workspace time zone and inbox hours.
4. Apply the policy to inboxes, or use automation to **apply SLA** when conditions match (for example a VIP label).
If several policies hit one conversation, follow the product’s stacking rules; prefer one main policy per inbox and override VIP via automation.
—
## 2. What agents see
Lists / detail may show FRT / NRT / resolution state (in progress, met, breached). Breaches can create “SLA breach” notifications.
Whether a bot’s first reply counts as first response depends on the current SLA definition. If performance reviews only count humans, do not treat bot FRT as an agent win.
—
## 3. Reports
**Reports → SLA** shows attainment and can export CSV. Use it for weekly reviews, not as the only basis for penalties — combine with inbound peaks and staffing.
—
## 4. Frequently asked questions
**Instant breaches?** Thresholds too short, or calendar time while you only staff daytime. Align with business hours.
**Policy never applied?** Inbox not linked, or automation conditions never matched. Old conversations are usually not backfilled; timing starts from events after apply.
—
## Related articles
– [Conversation workspace](.3-conversation-workbench)
– [Reports](./10-reports)
– [Account settings and security](./27-account-settings-and-security)