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

# Autopilot Settings: Language, Timing, and Escalation

> The workspace-wide autopilot settings applied to every recovery case: language and timezone detection, working-day sending, payment-failure consequences, and escalation rules.

These settings apply to **every** recovery case the autopilot runs. You set them once under **Auto-Pilot → Configuration**; the strategy algorithm and executor honor them on each case. You do not build flows here — you set the boundaries the automation operates within.

***

## Language & timezone detection

By default, Airdun reads each customer's country and language from their Stripe data, then sends every touchpoint during that customer's local working hours, in their language.

* **Auto-detection on** — country and language come from the customer's Stripe record. This is the recommended setting.
* **Fallback language and timezone** — used only when Stripe has no location or language for a customer. Set these to the values that best fit the majority of your customers.

<Note>
  Auto-translation is per customer and automatic — a customer in Germany receives a message in German with no manual setup. See [Brand Voice](/features/brand-voice) for how translation preserves your tone.
</Note>

***

## Send only on workdays

When enabled, touchpoints are paused on weekends and resume on the next workday. Use it to keep recovery messages inside a business rhythm; leave it off to let the algorithm reach customers whenever it predicts the best response.

***

## Payment failure consequences

Tell the model what actually happens to a customer when their payment keeps failing — for example, *active campaigns are paused* or *the account loses access after a grace period*. The strategy uses this to explain the stakes, in the customer's own context, as a sequence escalates.

* **Consequence description** — be specific. Vague copy produces vague messages.
* **Delay before consequence** — how many days of failed payment pass before that consequence applies. The model references this timeline when it sets expectations with the customer.

<Info>
  This only describes your product's own policy so messages can reference it. Airdun never changes anything in your product — it does not pause accounts or revoke access.
</Info>

***

## Escalation rules

An escalation rule flags a matching case to your team instead of running the standard automated sequence. When a new case matches a rule — or looks like a fraudulent decline — Airdun notifies your team in Slack and the in-app bell and hands the case over.

* Conditions **within** a rule are joined with **AND**.
* **Multiple** rules behave as **OR** — a case escalates if it matches any one of them.

For how notifications and Slack are wired, see [Notifications](/configuration/notifications).

<Note>
  You do not configure the recovery sequence itself — Airdun designs it per case. What you control here is language, timing, the consequence narrative, and which cases skip automation for a human. See [Recovery Strategy](/how-it-works/recovery-strategy).
</Note>