This document describes how to use the workflow canvas to create a workflow using either natural prompts or the visual editor.
Before you begin
Before using this feature, complete the following steps:
- Ensure that you have a web app with the necessary data sources connected. For more information, see Create an app, Connect a Google data source, and Connect a third-party data source.
- Ensure that you have enabled the Workflow Builder feature toggle for your web app. For more information, see Manage web app features.
Create a workflow using natural prompts
App
To create a workflow agent using natural language prompts, you describe the task you want the agent to perform, and Workflow Builder generates a workflow scaffold for you. For more information, see Create a workflow using natural language.
Create a workflow using the visual editor
App
To create a workflow using the visual editor, do the following:
- Open your Gemini Enterprise web app in a browser.
- In the web app navigation menu, click + New agent > Workflow.
- Click Build manually to open the visual editor.
- A default manual trigger is added to the canvas. Click Input fields to enter a required or optional input field that the user can provide when
running the workflow.
To change the trigger to another trigger type, click the manual trigger in the canvas and then select another trigger type in the configuration pane.
- Click + Add step to add additional steps to the workflow. For the list of supported steps and their configurations, see Steps.
- To preview and test your workflow, switch to the Test tab and start a test run.
To turn on your workflow agent and make it live, click Turn on.
The workflow agent is live and starts to execute according to the configured schedule or events, or manually by the user.
What's next
- Learn how to run test simulations for your workflow.