Intempt Docs
GuidesWorkflows

Workflows

Workflows is a canvas for automating what happens to your data: a trigger starts a run, and the nodes after it act on the record that started it.

Overview

Workflows is a visual builder for automating what happens to your data. You draw a graph: a trigger node at the start, one or more action nodes after it, and connections between them. When the trigger fires, Intempt executes the graph once per matching record and records the result as a run.

A workflow is edited on a canvas, published as a version, and then monitored on its own Runs tab. Nothing runs until you publish, and a published workflow keeps running the version you published even while you edit a new draft over it.

A live workflow on the Build canvas, with the header bar, Add node, the toolbar and two connected nodes

Workflows and Journeys

Both products use a node canvas, and they solve different problems:

  • Journeys orchestrate messaging to people. You enroll users into a flow and send email, SMS, push and in-app messages with waits and branches between them. See Journeys.
  • Workflows move and act on data. The nodes available today start a run from an event, a raw event or a schedule and hand the result to a system outside Intempt.

If your goal is to reach a person, build a journey. If your goal is to get data somewhere or act on a record, build a workflow.

What you can build today

Workflows ships the canvas, the publishing model, the version history and the run monitoring in full. The node catalog is being filled in over time, so the node picker lists every node the product intends to offer and only lets you place the ones that are built.

Five nodes are available right now:

NodeCategoryWhat it does
Event triggerStartStarts a run when a matching event is recorded or an attribute changes
Scheduled triggerStartStarts a run on a repeating schedule
Raw events triggerStartStarts a run on every raw event from the Events only sources you pick
Publish to KafkaSendPublishes each record to a Kafka topic as a key and a value you author
Write Parquet to S3SendWrites events, user records or failed events to Amazon S3 as Parquet files or an Iceberg table

Every other row in the node picker is visible but greyed out, with the hint "Not available yet - this node has not been built." Those rows become selectable as the nodes ship; nothing about your existing workflows changes when one does.

📘 Good to know

A workflow needs at least one trigger and at least one action before it can be published, and every node on the canvas has to be connected to the rest of the graph. An unconnected node never runs.

Finding your workflows

Click Workflows in the left navigation to open the workflow list. The table shows one row per workflow with these columns:

  • Name - the workflow's name.
  • Status - whether the workflow is a draft, running or stopped.
  • Runs - how many times this workflow has run.
  • Success rate - the share of runs that completed successfully.
  • Last updated - who last changed it, and when.
  • Created by - the team member who originally created it.

Click Create workflow to start a new one, or click any row to open it. The row's Actions menu has a Delete option.

The Workflows list table showing several workflows with their status, runs and success rate, and the Create workflow button

Workflow status

The status badge sits next to the workflow's name in both the list and the builder header.

StatusWhat it means
DraftNever published. The canvas is the whole workflow and nothing is running.
LivePublished and taking new runs.
PausedPublished, but not taking new runs. Runs already in flight are held until you resume.
StoppedNot taking runs and not resumable by publishing. Resume it before publishing changes.
ScheduledNot yet published, and treated the same way a draft is.

Where to start

On this page