01Who this is for
- Founders who want to confirm AI crafted automations operate correctly before customers see them.
- Marketing agencies building client automations that need a quick preview of tag, note, and messaging actions.
- SaaS product teams with complex triggers that require in chat visibility of each step’s outcome.
02AI Builder workflow testing, explained
AI Builder testing runs an automation that was created or edited by the AI directly from the chat window. When you start a test, the platform performs the real actions defined in the workflow, adding tags, creating notes, sending messages, so you can see exactly what will happen.
During execution, each action on the canvas displays its current state, such as Running, Success, or Action skipped. After the run finishes, an AI generated summary lists which steps succeeded, were omitted, or produced errors, and you can click any action to view its input and output data.
The feature lives inside the AI Builder chat and is optional; you may choose to publish a workflow without testing, but testing is recommended to catch configuration problems early.
03How businesses use AI Builder workflow testing
- Dental practice. Select a test patient as the contact, run the workflow through AI Builder, and watch the canvas to confirm that appointment reminders and follow up messages are sent correctly. Inspect each step’s input and output to verify data handling.
- Coaching business. Enroll a sample client, start a test run, and monitor the sequence that delivers onboarding emails and assigns tags. Review the AI summary to ensure no actions are unintentionally skipped.
- Ecommerce store. Use the test using events option to simulate a purchase event, then observe how the workflow creates order notes and triggers confirmation texts. Adjust the flow and run another test until the summary shows all actions succeeding.
04Before you start
- The New Test Workflow must be activated in HighLevel Labs.
- A test contact should exist to avoid affecting real customers.
- AI Builder must have completed a workflow creation or edit request before the test button appears.
05How to set up and use AI Builder workflow testing in HighLevel
Run a workflow test using AI Builder
- Open HighLevel Labs and enable the feature called The New Test Workflow.
- Navigate to Automation > Workflows and open the desired workflow.
- Launch AI Builder and ask it to create or modify the workflow.
- When AI Builder finishes, click the Test workflow button inside the chat.
- Choose the Enroll a contact testing method.
- Select a contact from the list or review the presented test data.
- Press Start test run to begin watching the workflow on the canvas.
- Read the AI generated summary and click any action to see its input and output.
- If changes are needed, edit the workflow and click Run another test to repeat.
06Settings that matter
- The New Test Workflow. Turns on the ability to launch and monitor workflow tests from AI Builder.
07Things to know before you rely on it
- Tests run real actions such as adding tags, creating notes, and sending messages; use a test contact to keep real customers unaffected.
- Running a test is optional; you can still save and publish a workflow without testing.
- A successful test does not guarantee the workflow will behave identically for all contacts or trigger conditions.
08If something goes wrong
- Why was an action skipped during testing?. Open the AI generated summary after the run and look for the skipped step to understand why it was omitted.
- I cannot see the Test workflow option.. Verify that The New Test Workflow is enabled in HighLevel Labs and that AI Builder has finished a create or edit request.
09How it connects to the rest of HighLevel
- Explore how to build automations with AI Builder AI Builder overview
- Get started with workflow creation and basic triggers Workflow basics
- Review all workflow features and options All Workflows features
Michael's take

Skipping testing can let tiny bugs slip into a live automation, and because tests fire real actions, they can clutter your tags or notes if you use a real customer by mistake. The risk of surprise messages or data changes is real.
When you run a test inside AI Builder, you get a quick visual of each step and a concise summary that tells you exactly what succeeded or why something was skipped. That instant feedback is worth the extra few clicks before you push a workflow live.