# Automations

> Learn how to build, save, and turn on automations.

Source: https://www.lev.com/docs/learn/originate/automations

Last updated: 2026-09-25

---
Automations lets you run a saved workflow when a schedule fires, a pipeline record changes, or a campaign enrollment replies, bounces, or exits. The builder is a full page under *Automations* in the sidebar, when that item is available on your account. You add a trigger and at least one step, save a version, and turn the automation on.

## Create an automation {#create-an-automation}

- **Open the list** — Select *Automations* in the sidebar, then *New automation*.
- **Start the automation** — In *Create a new automation*, describe the workflow under *Build with Lev Agent*, choose *Start from scratch*, or pick a template. Categories are *Suggested*, *Email*, *Pipeline*, *Campaign*, and *Scheduled*.
- **Name it** — A from-scratch automation opens as Untitled automation. Edit the name in the builder header.
- **Rename from the list** — Choose *Rename* to open *Rename automation*.
- **Archive it** — Choose *Archive* on the row.
- **Show archived rows** — Turn on *Show archived* to see archived automations and restore one.

Choosing a template names the new automation. Add the trigger and steps on the canvas. An empty canvas cannot be saved until both are in place.

## Build the workflow {#build-the-workflow}

The canvas is the automation. A trigger decides when it runs, and each step is what it does.

- **Add a trigger** — Choose *Add trigger*, or *Add a trigger* on an empty canvas. A schedule uses *Frequency* (*Hourly*, *Daily*, *Weekly*, or *Monthly*), plus *Day*, *Day of month*, *Time*, and *Timezone* when that frequency needs them.
- **Scope a reply or bounce** — Point the trigger at a campaign.
- **Scope an exit** — An enrollment-exited trigger uses *Any campaign* or one campaign, and *Any reason* or a reason: *Engaged*, *Exhausted*, *Declined*, *Bounced*, *Canceled*, or *Relaunched*.
- **Add a step** — Choose *Add step*. Steps include *Send a reply*, *Draft a reply*, *Enroll in a campaign*, *Update contact in pipeline*, and *Remove contact from pipeline*.
- **Set who writes a reply** — On *Written by*, choose *Lev writes it* or *I write it myself*. *Draft a reply* saves the message for you to send. *Send a reply* sends it.
- **Match a message** — On a trigger that supports it, use *Read message for* to describe what the message should match before the run starts.
- **Turn it on** — Choose *Turn on*. The control reads *Live* while that version is on, and *Off* when it is not. Turn it off with *Turn off*.

Pipeline triggers run when a record is added to a pipeline or updated in one. Schedule a Marketing status report from this builder when that agent step is available on your account.

## Build with AI {#build-with-ai}

*Build with AI* drafts the workflow from a description, on the builder page.

- **Describe the workflow** — In *Create a new automation*, write the description under *Build with Lev Agent* and send it. The builder opens with that prompt.
- **Revise on the canvas** — Choose *Build with AI* in the builder header to generate or adjust the trigger and steps.
- **Review before you turn it on** — Check the trigger and each step, then save. Nothing runs until the automation is *Live*.

## Versions {#versions}

*Version history* is the list of saved versions. Promoting a version is what makes it the live one.

- **Open version history** — In the builder header, choose *Version history*.
- **Promote a saved version** — When the version differs from the live one and you have no unsaved edits, choose *Promote*.
- **Save edits onto the live version** — Choose *Save and promote* to save the canvas and make that version live.
- **Keep edits as a draft** — Choose *Save as draft* when you want the changes stored without replacing the live version.
- **Return to the canvas to change Live** — While you are viewing history, turn the automation on or off from the canvas.

An automation saves only when you save it. *Save and promote* and *Save as draft* are the save actions. The header badge *Per run* shows the credit charge reserved for each run, when that automation bills credits.

## Run history {#run-history}

*Run history* lists past runs and the steps each run took.

- **Open run history** — In the builder header, choose *Run history*.
- **Filter by status** — A run can be *Succeeded*, *Failed*, *Running*, or *Skipped*.
- **Read the credit charge** — The selected run shows the credits charged for that run. A step that billed credits shows its own charge. A message condition shows *Condition credits* and *Each time this condition is checked* when that check bills credits.