Skip to main content
The automation editor gives you a visual canvas where you connect nodes to build multi-step workflows. Every workflow starts with a Trigger node and flows through any combination of Action, Condition, and AI Step nodes.

Opening the Editor

  1. Go to your agent’s sidebar and click Automations
  2. Click New Automation to open a blank canvas, or click an existing rule to edit it
  3. Give your automation a name and click into the canvas to start building

Node Types

Trigger

The starting point of every workflow. Set it to Conversation Started, Product Inquiry, Order Inquiry, or Custom Keyword.

Action

Performs an operation: send a reply, add to a Mailchimp list, fire a webhook, send an email, and more.

Condition (If/Else)

Evaluates a variable and routes the workflow down one of two branches depending on whether the condition is true or false.

AI Step

Calls the AI with a custom prompt. The output is stored in a named variable you can use in downstream nodes.

Creating a Workflow

1

Set Your Trigger

Click the Trigger node on the canvas and choose when this automation should fire.For Custom Keyword, enter one or more keywords and choose Any (fires if any keyword matches) or All (every keyword must appear in the message).
2

Add an Action Node

Click the + button that appears on the edge of any node, or drag a new node from the sidebar panel onto the canvas.Connect nodes by dragging from one node’s output handle to another node’s input handle.
3

Configure the Action

Click any Action node to open its configuration panel on the right. Select the action type and fill in the required fields.See Action Types below for all available options.
4

Add Conditions (Optional)

Insert a Condition node between two Action nodes to branch the workflow based on customer data.Each condition evaluates a variable (e.g., customerEmail, channelType) using an operator, then routes to either the True or False branch.
5

Test and Save

Click Test to simulate the automation against a sample conversation. Review the output, then click Save to activate.Toggle the Active switch on the automation list page to pause or resume a rule at any time.

Action Types

Reply

Send a message back to the customer directly through the chat. Reply actions run before the AI responds — they replace the AI reply entirely.
OptionDescription
TextA plain-text message. Supports {{variables}} for personalisation.
FileShare a file by entering its URL and display name.
FormDisplay a dynamic form with custom fields to collect customer information.
When a customer submits a form, the workflow automatically resumes and maps form values as {{form_fieldname}} variables for use in downstream nodes.

Email

Send an email notification to any address when the automation fires.
FieldDescription
ToRecipient email address. Supports {{variables}}.
SubjectEmail subject line.
BodyEmail body template. Supports {{variables}} and markdown.
Reply ToOptional reply-to address.
CC / BCCOptional additional recipients.

Webhook

Send conversation data to an external URL via HTTP.
FieldDescription
URLThe endpoint to call.
MethodGET, POST, PUT, or DELETE.
AuthNone, Bearer Token, API Key, or Basic Auth.
Body TemplateJSON body with {{variable}} placeholders.
The webhook response is stored in {{webhook_response}} and available to downstream nodes.

Mailchimp

Add or remove a subscriber from your Mailchimp audience, or manage their tags. Requires a Mailchimp integration to be connected first.
Sub-actionWhat it does
add_to_listSubscribes the customer to a selected audience
remove_from_listArchives the subscriber from the audience
add_tagApplies a tag to the subscriber
remove_tagRemoves a tag from the subscriber

MailerLite

Add or remove a subscriber from your MailerLite group, or manage their tags. Requires a MailerLite integration to be connected first.
Sub-actionWhat it does
add_to_groupAdds the customer to a selected subscriber group
remove_from_groupRemoves the customer from the group
add_tagApplies a tag to the subscriber
remove_tagRemoves a tag from the subscriber

AI Step

Run a custom AI prompt mid-workflow and store the result in a variable.
FieldDescription
Prompt TemplateYour instruction to the AI. Supports {{variables}}.
Output VariableThe name to store the result under (e.g., sentiment). Use it downstream as {{sentiment}}.

Available Variables

Use {{variable}} in any text field, body template, or prompt. ChatRos automatically populates these from the current conversation:
VariableValue
{{customerName}}Customer’s display name
{{customerEmail}}Customer’s email address
{{customerPhone}}Customer’s phone number
{{conversationId}}Unique conversation ID
{{agentName}}Your agent’s name
{{workspaceName}}Your workspace name
{{channelType}}Channel (e.g., whatsapp, messenger, widget)
{{message}}The customer’s most recent message
{{country}}, {{city}}Geolocation data (if available)
{{timestamp}}Current date and time (ISO 8601)
{{form_fieldname}}Value from a submitted form field
{{outputVariable}}Output from an AI Step node

Condition Operators

Condition nodes support the following comparison operators:
OperatorDescription
equals / not_equalsExact match
contains / not_containsSubstring match
starts_with / ends_withPrefix or suffix match
greater_than / less_thanNumeric comparison
greater_than_or_equal / less_than_or_equalNumeric comparison with equality
is_empty / is_not_emptyChecks if the variable has a value
matches_regexTests against a regular expression

Testing Your Automation

Click the Test button while in the editor to simulate a conversation event. ChatRos will walk through your node graph and show you:
  • Which nodes were reached
  • What data each action would have sent
  • Any errors in template variables or missing configuration
Test mode does not send real emails, call live webhooks, or modify Mailchimp/MailerLite subscribers. It is a dry run only.

What’s Next?

Template Gallery

Start from a pre-built template instead of building from scratch

Mailchimp Integration

Connect your Mailchimp account

MailerLite Integration

Connect your MailerLite account

Automations Overview

Back to the Automations overview