Creating Workflows

This guide walks you through creating different types of workflows to automate actions in your AI support system.

Creating a Workflow

1. Go to AI Agents > Workflows
2. Click Create Workflow (or New)
3. Fill in basic information
4. Configure the action type
5. Set up triggers
6. Save and test

Basic Information

Name

Give your workflow a descriptive name:

• "Cancel Order Request"
• "Collect Lead Info"
• "VIP Customer Alert"

Description

Describe when this workflow should trigger:

Triggers when a customer wants to cancel their order. 
Sends order ID and customer email to our order management system.

This description helps the AI understand when to use the workflow.

Slug

A unique identifier used in the system:

• cancel-order
• collect-lead
• vip-alert

Automatically generated from name if not specified.

Webhook Workflows

Send data to any HTTP endpoint.

Configuration

Example: Order Cancellation

URL: https://api.yourstore.com/orders/cancel
Method: POST
Headers:
  Authorization: Bearer your-api-key
  Content-Type: application/json
Payload:
  order_id: {{order_id}}
  customer_email: {{customer_email}}
  reason: {{cancellation_reason}}

Available Variables

In your payload, use these variables:

Testing Webhooks

1. Use a tool like webhook.site to test
2. Configure your workflow to send there
3. Trigger in preview mode
4. Verify data received

Zapier Workflows

Connect to Zapier for no-code integrations.

Setup

1. Select Zapier as action type
2. Click Connect to Zapier
3. Authorize the connection
4. Select or create a Zap

Common Zaps

• Google Sheets: Log conversation data
• Slack: Alert team of escalations
• HubSpot: Create contacts from leads
• Trello: Create cards for follow-ups

Make (Integromat) Workflows

Connect to Make for advanced automations.

Setup

1. Select Make as action type
2. Click Connect to Make
3. Authorize the connection
4. Select your scenario

Advanced Features

Make supports:

• Multi-step scenarios
• Conditional branching
• Data transformation
• Error handling

Escalation Workflows

Hand conversations to human support.

Configuration

Example Configuration

Email: support@company.com
Include Transcript: Yes
Subject: [Escalation] {{topic}}
Collect Fields:
  - Email (required)
  - Phone (optional)
  - Preferred callback time

Tag Workflows

Automatically apply tags to conversations.

Configuration

Example

Tag: "Billing Issue"
Condition: Intent matches billing-related queries

Custom Reply Workflows

Send specific responses when triggered.

Configuration

Example

Message: "I've initiated your cancellation. You'll receive a confirmation 
email within 24 hours. Is there anything else I can help with?"

Persona Scoping

Limit workflows to specific personas:

1. In workflow settings, find Persona Scope
2. Choose:
   • All - Works with all personas
   • Selected - Choose specific personas
3. Select personas if limited

Enabling/Disabling

Toggle workflows on/off:

1. Go to AI Agents > Workflows
2. Find your workflow
3. Toggle the Enabled switch

Disabled workflows won't trigger but retain configuration.

Testing Workflows

Before going live:

1. Go to Widget > Preview
2. Start a conversation
3. Trigger the workflow condition
4. Check execution in workflow history
5. Verify external system received data

Best Practices

Start with Description

Write a clear description before configuration. This helps the AI know when to trigger.

Use Test Endpoints First

Don't test against production systems. Use:

• webhook.site
• requestbin.com
• Staging/test endpoints

Handle Errors Gracefully

Configure fallback messages:

If workflow fails, respond: "I've noted your request and our team 
will follow up within 24 hours."

Monitor Initially

After launch, watch:

• Execution counts
• Error rates
• Response times