Automation and AI

Build, test, and activate a workflow

Create a Workspace369 automation with a precise trigger, filters, sequence actions, a simulated test run, and a reviewable live run history.

View as Markdown

Build a workflow as a draft, test the exact route with controlled data, and activate it only after the trigger, audience, timing, and delivery content are ready.

Safe publishing sequence: Draft → controlled test → trace review → activation → live-run monitoring.

Start the workflow

From Automations, start a blank workflow or use an available draft starting point. Give it a name that describes the event and outcome, such as New lead follow-up rather than Automation 3.

The builder separates the workflow into three parts:

  1. Trigger and filters.
  2. Sequence actions.
  3. Review run.

Choose the trigger and enrollment

Select the event that should start the workflow. Available triggers depend on the feature area and can use structured client-work events and statuses.

Workspace369 workflow editor for a new client welcome sequence with a Client Created trigger set to run once per client, two email steps separated by a one-day wait, a Delivery check marked Ready, and the Update workflow button
The Edit tab lays out the trigger, each step, and the waits between them as one path. The Delivery check on the right must read Ready before the workflow can deliver anything.Workspace369 2026.3 · iPad · Light mode Open full size

Enrollment controls whether the same record can enter repeatedly. When the builder shows Run once per record, the workflow prevents duplicate runs for that record. Confirm that behavior matches the process before activation.

Add filters

Filters decide which triggered records qualify. Use the smallest set that clearly represents the audience.

Review a filter when a status, category, task board, client type, priority, location, or other structured value changes. A renamed or retired value can stop the intended records from qualifying.

Configure sequence actions

Available follow-up actions include structured Inbox email, SMS, and internal notification steps. Additional steps can be added to form a sequence, with a wait between steps when the follow-up should not happen immediately.

For every step, verify:

  • The channel and recipient.
  • The subject or message content.
  • Variables rendered from the triggering record.
  • Step-specific filters.
  • The wait before the next step.
  • The person responsible when the action needs review.

Keep a workflow in Draft while any of those details are uncertain.

Save and run a test

  1. Save the workflow while it is still in Draft.
  2. Open the run ledger and choose Test run.
  3. Select controlled data that represents the intended audience without using a live client action as an experiment.
  4. Run the simulation and open its trace.
  5. Keep the workflow in Draft until the route, rendered content, wait steps, and exit behavior are correct.

Test runs are simulated and identified separately from live executions. Their status language describes what the workflow would do; it is not proof that a real client received a message.

Review the test trace

Open the test run and inspect:

  1. The record used for enrollment.
  2. Trigger and filter evaluation.
  3. The sequence of steps.
  4. Rendered subject and message content.
  5. Any failed or skipped step.
  6. The final exit from the workflow.

Correct the draft and run another controlled test when the route is not what you expected.

Activate the workflow

Activate only after someone responsible for the process has reviewed the test. Activation allows qualifying live events to enter the workflow.

Changing an active workflow can affect future runs. Review enrollment, timing, audience, and client-facing copy again before saving a material change.

Monitor live runs

Use View runs or the run ledger to review live and test activity. Run status can distinguish delivered, failed, and pending work. Open a run for its route, timing, rendered content, and error detail.

Retry a failed action only after correcting the cause and confirming that a retry will not duplicate a message or record change.

Return to Automations for workflow planning and governance guidance.

Was this article useful?

Your answer helps us decide what to clarify next.