This section showcases how to combine Canvas features like global nodes, dynamic conditions, and rewind levels to create powerful, dynamic chatbot flows.

​
Example 1: Multi-Step Booking

Scenario: A chatbot for scheduling appointments that handles errors and provides user-friendly navigation.

​
Flow Overview:

  1. Start Node: Greets the user and collects their initial details.
  2. Dynamic Conditions: Routes users to different booking steps based on input.
  3. Global Node: Uses a reusable “Cancel” node to handle cancellations at any step.
  4. Rewind Levels: Allows users to go back and correct inputs if needed.
How It Works:
  • If the user provides invalid details, the chatbot rewinds to the input step.
  • If the user says “Cancel,” the chatbot routes to the global Cancel Node.

​
Example 2: Personalized Recommendations

Scenario: An e-commerce chatbot providing tailored product recommendations.

​
Flow Overview:

  1. Start Node: Retrieves user preferences using global variables (e.g., user_name, shopping_category).
  2. Dynamic Conditions: Suggests products based on user preferences and input.
  3. Global Node: Provides a reusable Help Node for FAQs.
  4. End Node: Concludes with a personalized thank-you message.
How It Works:
  • The chatbot dynamically suggests products by matching user input with a product database.
  • Users can ask for help or clarification at any time using the Help Node.

​
Example 3: Error Handling with Rewind

Scenario: A support chatbot that manages errors and retries while preserving context.

​
Flow Overview:

  1. Start Node: Welcomes the user and asks for their issue.
  2. Default Node: Processes the issue and attempts to resolve it.
  3. Rewind Levels: Allows retries if the tool call or logic fails.
  4. Global Node: Routes escalated issues to a human agent or advanced support system.
How It Works:
  • If the initial resolution fails, the chatbot rewinds to the user input step and requests clarification.
  • Escalation is handled through a global node for consistent fallback behavior.

​
Best Practices for Advanced Flows

  • Combine Features: Use global nodes, dynamic conditions, and rewind levels together to maximize flexibility.
  • Test Edge Cases: Simulate scenarios with invalid inputs, tool failures, and unmatched conditions.
  • Leverage Personalization: Use global variables to create tailored experiences for users.
  • Simplify Logic: Avoid overcomplicating flows; prioritize clarity and maintainability.

​
Example Flow Testing

After designing advanced flows, test them thoroughly:
  1. Simulate user inputs and follow the flow progression.
  2. Test dynamic conditions to verify proper routing.
  3. Validate rewind functionality to ensure context is preserved.
Ready to apply these techniques? Return to the Canvas Workspace to implement these advanced flow designs in your chatbot!