Actions
Make your No-Code Actions available to the Agent so it can perform automated tasks for your users during conversations. When a user asks to do something, it can run the matching Action automatically.
Overview
The Actions tab lists every No-Code Action you have connected to the Agent. Each Action card shows its name, description, step thumbnails, and status. From here you can activate or deactivate each Action for the Agent, configure its trigger instructions, or remove it.
Dashboard path: Agent > Knowledge > Actions
Prerequisite: Actions must first be created in the Automated Actions section of the dashboard. Only existing No-Code Actions can be added to the Agent.
Adding an Action
Click + Add Actions (top-right) to open the Action selection modal.
Browse or search your existing No-Code Actions. Each card shows step thumbnails, the Action name, description, run count, and last updated date. Click an Action to select it.
After selecting an Action, the Instructions modal opens to configure how the Agent should use it.
Configuring Action Instructions
The Instructions modal defines when and where the Agent should run an Action. The workflow mirrors the Tour Instructions pattern with the same two-step process: Topics first, then location.
When should Agent run this action? (Topics)
Define the intents, questions, or phrases that should trigger this Action.
Adding Topics:
Click + Add Topic to type a custom intent or question.
Each topic represents a user intent (e.g. "Open the surveys section", "How do I invite a team member?", "Navigate to settings").
Add as many topics as needed to cover different phrasings.
Suggestions from Jimo AI:
Just like with Tours, Jimo AI generates topic suggestions based on the Action's content and steps. Click + to add any suggestion instantly.
Where should Agent run this action?
Controls on which pages and environments the Agent is allowed to run this Action.
Environments: Choose which deployment environment(s) this Action applies to (e.g. "Any domain" or specific environments).
Page rules: Define URL-based rules for where this Action can run.
+ Add rule: Add a path condition manually.
✨ Generate rules: Let Jimo AI suggest rules based on the Action's steps.
Fallback URL (optional): A URL to redirect users to if no page rules match.
Click Save Trigger to save the configuration.
Menu options
The three-dot menu (⋮) in the top-right of the Instructions modal provides additional options:
Test: Run the Action in a test environment to verify it works as expected.
Redo: Re-record the Action steps if the UI has changed.
Duplicate: Create a copy of this Action configuration.
Delete: Remove this Action from the Agent.
Managing Actions
From the Actions list, each Action has the following controls:
Active / Inactive toggle: Enable or disable the Action for the Agent.
Instructions button: Open the Instructions modal to edit topics, location, and trigger settings.
🗑 Delete: Remove the Action from the Agent entirely.
How it works at runtime
When a user sends a message in the Agent chat:
The Agent analyzes the user's intent.
It compares the message against all active Action topics.
If a match is found and the user is on a qualifying page, the Agent runs the No-Code Action: it takes control of the UI and performs the recorded steps automatically (clicking buttons, navigating pages, filling fields).
If the user is not on a qualifying page, the Agent can use the Fallback URL to redirect the user first.
The Action executes as a standard No-Code Action, performing each recorded step sequentially. The user sees the UI responding in real-time as the Agent works.
Best practices
Last updated