Automate repeat production

Automations connect an input to a sequence of production steps and a destination. Use them after the product model, template, and expected output have already succeeded in a controlled run.

Automation removes repeated handoffs. It does not remove the need for a known-good setup, observable run state, and visual review.

Start with a proven path

Before building an automation:

1. Prepare a representative product model.
2. Produce an accepted result from the intended template.
3. Confirm every required angle or clip.
4. Test the destination with one output.
5. Write down the failure that should stop the workflow.

The first automated run should reproduce work you understand. Do not use an automation to discover the creative setup and production rules at the same time.

Choose how work begins

• Manual: Run a saved workflow when you choose.
• Watch Folder: Detect supported files in Google Drive, OneDrive, or Adobe Experience Manager.
• Model Ready: Start after a Glossi model finishes processing.
• Project Created: Start from a new project.
• Render Complete: Continue from a finished render, often into a style, edit, or export stage.

Configure triggers

Build the production path

An automation is a directed graph. Each node receives the result of the previous stage and performs one supported job.

A dependable first workflow is:

Watch Folder -> Render from Template -> Export

Once that path is verified, add a branch or controlled adaptation step:

Watch Folder -> Render from Template -> AI Style -> Export

The core shipped path supports connector intake, template-based rendering, branching, supported style or AI-edit stages, and connector export. Material-swap and direct-colorway automation paths are still partial. Keep those steps out of unattended production unless the exact workflow has been validated with the Glossi team.

Understand automation nodes

Configure every dependency

The editor identifies incomplete nodes before a run starts. Check:

• source account and folder
• supported input types
• template and template version
• required angles or clips
• style or edit inputs
• destination account and folder
• branch-specific output behavior

Save the automation only after each node represents a real production decision.

Test before enabling the trigger

Use one representative source file. Follow it through the run and confirm:

• the correct input was detected
• the expected product model was created or selected
• the intended template version was used
• every requested output completed
• any derivative remained acceptable
• the correct files reached the destination

Folder triggers poll on a schedule, so they can add minutes before work begins. They are not a real-time input channel.

Operate from runs, not assumptions

The Runs view is the record of what happened. Each run exposes supported stage state, outputs, and failures. Use it to isolate the failed node, retry supported work, or cancel a request that should no longer continue.

Connector authorization can expire or be revoked. Restore access before retrying an affected intake or export stage.

Inspect runs and recover failures

Choose a different production surface when needed

Success state

An automation is ready when a new valid input can travel through the intended nodes, produce the expected outputs, arrive in the correct destination, and leave enough run history for an operator to understand a failure.