Run and Review Workflow Tests Inside HighLevel AI Builder

TL;DR

HighLevel added inline workflow testing directly inside the AI Builder chat, accessible via a Labs toggle.

You only need to act if you build automations using AI Builder and want to test runs without leaving the chat window.

Who this affects

HighLevel has not confirmed specific plan gating or sub-account restrictions for this feature.

It is relevant to anyone building or adjusting workflows through AI Builder who has access to HighLevel Labs.

What changed

HighLevel added several test and debugging tools directly into the AI Builder interface:

• AI Builder replies now provide a Test workflow button whenever you finish building or editing a workflow.
• Users can initiate tests inside the chat using either Enroll a contact or Test using events.
• The workflow canvas displays live step statuses throughout the test, labeled as Running, Success, Action skipped, or End.
• The chat log tracks steps with checkmarks as they complete.
• AI Builder generates an end-of-run summary detailing steps run, steps skipped, and any errors or missing data.
• Clicking on any canvas step shows the Input and Output data recorded during that specific run.
• A Run another test button enables quick retesting after making adjustments.
• Test runs execute actual actions, meaning they will apply real tags, create actual notes, and send real messages.
• Access requires enabling The New Test Workflow in HighLevel Labs.
• HighLevel has not confirmed rollout status or whether this replaces an existing testing interface.

Where to find it

To turn the feature on, go to HighLevel Labs > The New Test Workflow.

Once active, open Automation > Workflows > AI Builder. The test prompts and buttons will appear in the chat stream after building or modifying a workflow.

How agencies will use this

An operator asks AI Builder to create or edit a workflow inside a sub-account. Once the AI generates the flow, the operator clicks the Test workflow button in the chat panel and selects Enroll a contact or Test using events.

During execution, the operator watches the canvas highlight steps as Running, Success, Action skipped, or End, while the chat outputs a list of processed steps and missing data. If an issue occurs, the operator clicks the specific canvas step to examine its Input and Output payload, updates the workflow, and clicks Run another test.

Because these tests execute live actions—including writing notes, assigning tags, and dispatching messages—operators will need to run tests against designated internal test contacts rather than live leads.

Our take

This is a standard workflow update. It saves a few clicks by letting you review input and output data without leaving the AI interface.

Keep in mind that tests run real actions. If you trigger an outbound message or an action that writes notes and tags, it will actually fire. If you do not use AI Builder to generate automations, this update will not affect your operations.