Skip to content

SMS Workflows

Guide AI agents in SMS conversations with goals and desired outcomes for conversation flows.

Overview

SMS workflows help guide your AI agents on goals and desired outcomes for conversation flows. Define what the agent should achieve, provide context, and configure steps that guide agent behavior in SMS conversations.

Note: SMS Workflows are designed for agent guidance in conversations, NOT for bulk campaign automation or bulk sending.

Workflow Configuration

Goal (Required)

The workflow goal determines when this workflow matches incoming SMS messages. Goals help the AI agent understand the desired outcome for the conversation.

Available Goals:

  • General Inquiry: Matches all messages (catch-all)
  • Lead Capture: Matches messages containing keywords like "pricing", "quote", "interested", "budget"
  • Help Desk Support: Matches messages containing keywords like "support", "help", "issue", "problem"
  • Sales Inquiry: Matches sales-related questions
  • Appointment Booking: Matches appointment/scheduling requests

Context for AI Agent

Provide context about your business, services, pricing, or any information the AI should know when processing SMS messages. The agent will use this as reference and adapt it naturally - it's not a template to fill in.

Example Context:

You are a helpful sales assistant. When responding to inquiries:
- Thank them for their interest
- Ask what specific products or services they're interested in
- Gather their contact information (name, company, phone) if not already provided
- Mention that someone will follow up with detailed pricing
- Be professional and friendly

Response Template

Optional template for agent responses. The agent will adapt this naturally based on the conversation context.

Priority

Higher priority workflows are checked first. If an SMS message could match multiple workflows, the one with the highest priority wins.

Assigned Agent (Optional)

Select a specific AI agent to handle SMS messages matching this workflow. If not selected, the system will use routing rules or default SMS agent.

Workflow Steps

Workflow steps define what happens when an SMS message matches. The AI agent will follow these steps to handle the message.

Extract Information

Extract information from the conversation for use in later steps or for saving to the database.

Configuration:

  • Fields: Comma-separated list of fields to extract (e.g., name,email,phone,company)

Send Response

Guide the agent on how to respond to the message.

Configuration:

  • Template: Response guidance or template
  • The agent will adapt this naturally based on conversation context

Create Support Ticket

Create a support ticket when certain conditions are met.

Configuration:

  • Priority: Ticket priority level

Book Appointment

Guide the agent to help book an appointment.

Configuration:

  • Link: Calendar booking link (e.g., Calendly URL)

Save to Database

Save conversation data or extracted information to a database table.

Configuration:

  • Table/Collection Name: Database table to save to

Creating a Workflow

  1. Navigate to SMS Workflows page
  2. Click New Workflow
  3. Enter workflow name
  4. Select a goal (required)
  5. Add context for the AI agent
  6. Configure workflow steps
  7. Set priority (optional)
  8. Assign to specific agent (optional)
  9. Set workflow as active (optional)
  10. Save workflow

Workflow Matching

Workflows are matched to incoming SMS messages based on:

  1. Goal type: General inquiry matches all, others match based on keywords
  2. Priority: Higher priority workflows are checked first
  3. Active status: Only active workflows are considered

Integration with SMS Conversations

SMS Workflows integrate with the SMS Conversations API to guide AI agents:

  • Workflows provide context and goals to the agent
  • Steps guide agent behavior during conversations
  • The agent uses workflow guidance to achieve desired outcomes

Best Practices

  1. Define Clear Goals: Use specific goals that match your use cases
  2. Provide Rich Context: Give the agent enough information to handle conversations effectively
  3. Use Appropriate Priority: Set higher priority for urgent or important workflows
  4. Test Workflows: Test workflow matching with sample messages
  5. Monitor Performance: Review how workflows guide agent behavior in actual conversations

autoch.at Documentation