AI Trigger
AI Triggers are powerful connectors that allow you to integrate AI into chatbot scenarios and any external tools flexibly and automatically. Instead of responding with static text, AI Triggers enable action triggering, automatic data collection, and personalized user experiences.
Example: When a user says, "I want to shop", the AI understands the intent and – instead of just listing products – immediately triggers a sales scenario. A webform appears right inside Messenger, allowing the user to select products, quantity, fill in details, and place the order directly – all seamlessly within the conversation.
How to Create an AI Trigger
In the AI integration interface → scroll down to the “Trigger Management” section → click Add.
Enter the required information to create a trigger:
Trigger Name:
This is the identifier of the trigger you are creating. The name is used to search for and distinguish between different triggers.
Note: Choose a clear, meaningful name related to the action the trigger handles.
Examples: schedule_appointment, purchase, save_phone_number
Description Trigger:
This describes the function of the trigger so you or your collaborators understand its purpose.
Suggestion: Begin the description with the phrase “Allows the user to…” to clearly state the trigger’s role.
Examples:
- “Allows the user to submit information to book an appointment.”
- “Allows the user to select a product and proceed to purchase.”
Save data to Attribute:
This defines which data will be extracted from the user and saved to a specific variable (attribute) in the chatbot system.
Note:
- Attribute names should be meaningful, easy to understand, and reflect the type of data.
- Avoid unclear names like "x1", "AAAA", etc.
Example: When a user sends a message like “My phone number is 0909123456”, the trigger activates, extracts the phone number from the message, and saves it into the predefined attribute.
Transfer to Flow:
You can select the scenario that will run when this trigger is activated.
Output Message:
This is the content that the system will send back to the user when a trigger is activated. A typical case is when you don’t want AI to auto-reply but instead want to control the response manually to guide the user as intended.
You can use this section to:
- Send default messages.
- Guide the user on next actions.
- Connect to a support agent.
- Stop the bot or switch to another flow.
If the user sends:
- “I need to speak with a support agent”
The trigger will activate and the chatbot will:
- Send a message like:
“Thank you! Please call 0909 123 456 or click the button below to connect with a support agent.” - Pause the bot to wait for the agent, or switch to manual support mode
Here are some examples of AI Triggers:
Trigger Name: Schedule Appointment
Description Trigger: Allows users to schedule an appointment.
Transfer to Flow: Activates the “book appointment” scenario with a webform to collect related information like time, date, and purpose.
Trigger Name: Purchase
Description Trigger: Allows customers to make a purchase, place an order, or buy a product.
Transfer to Flow: Activates the "shopping" flow with products listed in a slide or sales page, including full details and pricing.
Trigger Name: Connect user to human
Description Trigger: Allows customers to connect directly with human agents.
Output Message: "A consultant will contact {{gender}} {{full_name}} shortly. Please leave a question or provide your phone number for easier communication."
Handover to Human The bot will temporarily pause for that user, waiting for a staff member to take over or switch to manual support mode.
Trigger Name: Save Phone Number
Description Trigger: Customers provide their phone number.
Save data to Attribute: Select the data to collect as "phone number" and save it to the attribute "phone_number".
Save to Google Sheets: When connected, all attributes triggered by the AI Trigger will be automatically saved to the designated Sheets.
Trigger Name: Weather Forecast
Description Trigger: Allows users to retrieve current weather information for a specific location.
Save data to Attribute: Select the data to collect as "city" and save it to the attribute "city".
Transfer to Flow: Activate the "Weather Forecast" flow, where you will connect to an external API to fetch the weather data for the collected city.
