Automated Actions (No-Code)
Record click-by-click workflows on your product and let Jimo replay them for your users automatically. No code required: just point, click, and save.
Overview
The Actions dashboard is organized into two tabs:
No-code: All your recorded Automated Actions. Each card shows step thumbnails, a name, description, and run count.
MCP (Coming Soon): A future integration with the Model Context Protocol for more advanced action capabilities.
Dashboard path: Actions > No-Code
What Are Automated Actions?
Automated Actions let you record a sequence of interactions on your product (clicking buttons, typing into fields, navigating between pages) and turn them into replayable workflows. Once saved, these actions can be triggered by the Agent, connected to experience CTAs, or run on demand.
Think of them as macros for your UI: instead of writing step-by-step instructions, you record the steps once and Jimo executes them for your users.
Creating an Action
Step 1: Set the starting page
Click + New Action (top-right). A modal asks where the flow begins.
Enter the URL of the page where the workflow starts (e.g. https://yourapp.com/settings), then click Open Page. Jimo opens the page in a new window and prepares for recording.
Step 2: Record the workflow
Once on the target page, a floating prompt appears at the bottom of the screen.
Click Start recording, then perform the exact steps you want to automate: click buttons, type into fields, press keyboard shortcuts, navigate between pages. Jimo captures every interaction.
Recording toolbar
While recording, a toolbar at the bottom of the screen shows your progress.
Status indicator
Shows "Ready to select!" when waiting for your next interaction, or "X selected" with the count of recorded steps.
⟳ Restart
Clears all recorded steps and starts over. A confirmation dialog asks "You've recorded X elements. Restart capture?" before proceeding.
⏸ Pause
Temporarily pause recording without losing your steps.
🗑 Delete
Remove the last recorded step.
Finish Recording
End the recording session and process the captured steps.
Step 3: Processing
After clicking Finish Recording, Jimo processes your captured interactions and generates the automated action.
This takes a few seconds. Once done, the Action editor opens with your recorded steps.
Action editor
The Action editor displays your recorded workflow as a vertical step flow. Each step represents one interaction you performed during recording.
Step flow (left panel)
The left panel shows all recorded steps connected by flow lines. Each step card displays:
Step type icon: Color-coded by type (green for clicks, purple for inputs, yellow for keyboard events).
Step title: Auto-generated description (e.g. "Click on 'Sign In'", "Input", "Keyboard Event").
Screenshot preview: A visual snapshot of the target element on the page.
Description: Human-readable explanation of what the step does (e.g. "Click the Sign in button to log in.", "Type 'Example' in the input field.").
Click any step to open its detail panel on the right side.
Step detail (right panel)
When you click a step, the right panel shows its properties and lets you edit the element targeting.
Input text step
Description: What this step does (e.g. "Type text into the input field.").
Target Element: A visual preview of the element on the page, with the CSS selector path shown below (e.g.
.flex > .relative > .pl-8\.75).Reselect element: Re-record the target element by clicking it on the live page.
Edit selector: Manually edit the CSS selector for advanced targeting.
Input: The text value to type into the field (e.g. "Example").
Click element step
Description: What this step does (e.g. "Click on the target element.").
Target Element: Two targeting methods:
Element text: Match the element by its visible text content (e.g. the button label).
CSS selector: Match the element by its CSS path (e.g.
form > .text-center > .inline-block).Switch to auto selection: Let Jimo automatically determine the best selector.
Action controls
At the bottom of the editor:
▶ Play: Run the action on the target page to test it. The steps execute sequentially, and you can watch the UI respond in real-time.
↻ Redo: Re-record the entire action from scratch. Opens the recording flow again.
✅ Save: Save the current version. If steps have been edited, a banner appears at the top: "This walkthrough has unpublished updates" with Discard updates and Publish updates buttons.
Action detail panel
Click any action card from the dashboard to open its full detail view.
The right panel shows:
Allow Agent to run this action
A toggle that controls whether the Agent can automatically run this action for your users. Click the chevron (>) to open the full Agent Trigger configuration with Topics, Where rules, and Fallback URL.
Statistics
Three KPI cards tracking the action's performance:
Runs
Total number of times this action has been executed.
Finished
Number of runs that completed all steps successfully.
Completion
Percentage of runs that finished without errors (Finished / Runs).
Metadata
Users Reached: Number of unique users who have triggered this action.
Last run: Timestamp of the most recent execution.
Last modified: When the action was last edited.
Created by: The team member who originally recorded the action.
How Actions are triggered
Currently, Automated Actions can be triggered in one way:
Via the Agent: Connect the action in Agent > Knowledge > Actions and configure Topics and Where rules. When a user's message matches a topic, the Agent runs the action automatically.
Coming soon : CTAs & Progress Triggers
Automated Actions will also be available as a new action type for experience CTAs and progress triggers, alongside the existing actions like Navigate To, Launch Experience, and Run JS.
This will let you trigger recorded workflows from any button or completion event in your experiences.
Testing an Action
You can test an action in two ways: using the ▶ Play button in the action editor, or using the Test option from the card's three-dot menu (⋮) on the dashboard.
When a test runs, a progress indicator appears at the bottom of the target page.
The indicator shows which step is currently executing (e.g. "1 of 5") and a Stop button to cancel mid-execution. Jimo takes control of the UI and performs each recorded step sequentially: navigating, clicking, typing, exactly as you recorded. Use this to verify the action works correctly before connecting it to the Agent.
Managing Actions
Card actions
Each action card on the dashboard has a three-dot menu (⋮) with four options.
Test
Run the action on the target page to verify it works correctly.
Redo
Re-record the action from scratch. Opens the recording flow with a new starting URL prompt.
Duplicate
Create a copy of this action with all its steps and settings.
Delete
Permanently remove this action.
Agent Trigger (from action detail)
From the action detail panel, clicking the Allow Agent to run this action chevron opens the Agent Trigger configuration. This is the same interface documented in Knowledge > Actions, with two steps:
Step 1: Topics (When should Agent run this action?)
Define intents and questions that should trigger this action. Jimo AI generates suggestions based on the action's steps. Click Continue to proceed.
Step 2: Where (Where should Agent run this action?)
Set the environment, page rules, and fallback URL. Click Save Trigger to activate.
The saved trigger is visible in read-only summary on the action detail panel.
Best practices
Last updated