Enabling a story for Workbench requires specific configuration. Here's what you need to set up.
Send to Story configuration requirements
For a story to work with Workbench, you need to enable Send to Story by turning on the toggle in the story, selecting either Workbench or Workbench and Send to Story, and configuring two actions:
Input action: Receives parameters from Workbench when the story is triggered.
Output action: Returns results to Workbench when the story completes.
These actions create a clear interface between Workbench and your story, defining what data goes in and what comes back out.
Configure the input action
The input action defines what parameters your story accepts from Workbench. When you configure it, you'll specify parameter names (for example, email_address or alert_id), parameter types (text, number, etc.), and whether each parameter is required or optional.
When Workbench triggers your story, it will determine what values to use for each parameter from the conversation context, or prompt you if needed.
Configure the output action
The output action defines what data your story returns to Workbench. This might include results from API calls, processed data, status messages, or error information. The output action ensures Workbench receives what it needs to continue the conversation or take further action.
✋ Try this: Configure a story for Workbench
Enable for Workbench or Workbench and Send to Story
When configuring a story, the Enable for toggle has three options:
Send to Story: Makes the story available in the Send to Story feature, which lets you trigger stories from other parts of Tines. Workbench can't access it.
Workbench: Makes the story available as a tool that only Workbench can trigger.
Workbench and Send to Story: Makes the story available in Workbench and in the Send to Story feature, which lets you trigger stories from other parts of Tines.
Choose the option that matches how you plan to use the story.
❗️Important
Require confirmation in Workbench
You can configure stories to require confirmation before they run, just as you can with templates. Enable this for stories that make changes to external systems, access sensitive data, perform actions that can't easily be undone, or have significant cost or resource implications.
When confirmation is required, Workbench shows you the story name and input parameters before executing.
Configure a timeout duration
You can set a timeout duration for stories used as Workbench tools. If the story hasn't completed within that time, Workbench will conclude the tool call gracefully and try an alternative approach rather than waiting indefinitely.
You can configure the timeout in two places: in the Send to Story settings panel within the story (which becomes the default for all Workbench calls to that story), or at the per-tool level in an AI Agent action (which overrides the story-level setting).
This is particularly useful for stories that call external APIs or run longer processes, where an unpredictable delay shouldn't stall an entire conversation.
Enable and disable stories
Within Workbench, you can enable or disable stories using the green or grey dot next to the story name. Disabling a story is useful when troubleshooting, when you want to temporarily restrict what Workbench can do, or when a story is under maintenance.