The Say node is designed to output a single message before moving to the next node in your conversation flow. It offers two primary methods of message generation: dynamic messaging using AI and static messaging.
Adding Context with Variables:
To add a variable, type "{" to open a suggestion list of available options. Use the syntax {variable_name} in your prompt or static message input.
Dynamic Message Using AI
This method leverages AI to generate context-aware responses. It's particularly useful when dealing with variable data, such as API responses.
Example Scenario: After an API call returns order information (status: "Delivered", delivery time, address), you can use the Say node to craft a personalized message.
Configuration:
Add the variable "last_api_response" to your prompt
Explain the API response schema in your instructions
Prompt the AI model to craft a complete message
Sample Prompt: "Using the order information in {last_api_response}, craft a friendly message to the customer about their delivery. Include the delivery address and time."
Sample Output: "Great news! Your order has been delivered. It arrived at 123 Main St. at 2:30 PM today. Is there anything else you'd like to know about your order?"
Dynamic Message Configuration Options:
Prompt: Enter your instructions for the AI here.
Model: Choose from GPT4o, GPT4o Mini, GPT3.5 (available on all plans), or GPT4 and GPT4-turbo (available on paid plans). Select based on your specific needs and the complexity of the task.
Temperature: Adjust this setting to control the AI's creativity. Higher values (0.7-1.0) increase creativity, while lower values (0.2-0.5) produce more deterministic and accurate responses.
Static Message
This method uses a pre-written message template with variable placeholders.
Example: "Great news! Your order has been {order_status}. It arrived at {delivery_address} at {delivery_time}. Is there anything else you'd like to know about your order?"
In this case, the variables (order_status, delivery_address, delivery_time) will be automatically filled with the appropriate values from your data.
Post-Message Flow:
After outputting the message (whether dynamic or static), the conversation will proceed to the next node in the flow. If it's the last node, the flow will exit.
Best Practices:
Use dynamic messages for complex, context-dependent responses.
Opt for static messages when you need consistent, predictable outputs.
Test different AI models and temperature settings to find the best fit for your use case.
Ensure your prompts for dynamic messages are clear and specific to get the best results.
Regularly review and update your static message templates to keep them relevant and engaging.
By effectively utilizing the Say node, you can create more natural, context-aware conversations that enhance user experience and provide personalized interactions.