Trigger types

Every trigger type has an optional priority field. If a user's intent matches the trigger settings of two or more flows, the flow with the highest priority intent will be launched.

regex

Trigger the flow when the user's text input matches the regular expression. We recommend you use pythex to test your regular expressions.

Example

Create a new flow and copy-paste the following code into it.

states:
    start:
        component: meya.text
        properties:
            text: |
                flow.value == {{ flow.value }}
                flow.my_num == {{ flow.my_num }}

In the Trigger settings, choose the regex trigger type and enter the following pattern: (?P<my_num>\b[1-9][0-9]?\b). Leave the other settings at their defaults and click OK, then Save.

This flow will be triggered when someone enters a number between 0 and 100, even if that number is embedded in a string. The user's entire input is stored in flow.value, while the regex group named my_num is stored in flow.my_num.

Test it in the test chat window:

keyword

Trigger the flow when the user's input exactly matches the keyword.

This trigger can be handy when testing and debugging your bot. For example, you may have hidden flows that print information pertinent only to developers. You can restrict access by using the keyword trigger type with a hard to guess keyword, like t3St1ng.

In general, forcing users to guess the exact word or phrase that will trigger their intended flow creates a bad user experience. Use Natural Language Understanding when possible.

dialogflow

Uses third party NLU service Dialogflow to determine the user's intent and trigger the flow if a matching intent is found.

🚧

This trigger requires a Dialogflow v2 agent and the Dialogflow integration enabled. If you are using a v1 agent, upgrade your agent or use the api_ai trigger.

Data available in the flow scope

You can access details regarding the matched intent within within the flow scope.

api_ai

Uses third party NLU service Dialogflow (formerly known as API.AI), to determine the user's intent and trigger the flow if a matching intent is found.

🚧

This trigger does not support Dialogflow v2 features, such as multi-language agents. If you wish to make use of those features, use the dialogflow trigger with your v2 agent.

Example

In DialogFlow, create a new agent and set the language to English. Create two intents, cancel and subscribe. Give each intent 3-5 training phrases and test them out.

In your agent's settings, note your client access token.

Back in Bot Studio, create a new flow and copy-paste the code below.

states:
    start:
        component: meya.text
        properties:
            text: You triggered this flow!

For the Trigger type, select api_ai and paste your client access token. Enter en as the agent's language. Leave the other fields with their default values, click OK, then Save.

Test the flow in the test chat window:

Note either Dialogflow intent will trigger the flow. This is because we didn't specify which Dialogflow intent we wanted to match. This is important to do since we may want to use the same NLU agent to trigger a cancel_subscription flow as well as a create_subscription flow.

Open the Trigger window again and enter cancel in the intent field. Click OK, then Save. Now test the the flow again.

This time, only the cancel intent will trigger the flow. The debug console confirms that start a new subscription didn't match.

Data available in the flow scope

You can access details regarding the matched intent within within the flow scope.

wit

Uses third party NLU service Wit.ai to determine the user's intent and trigger the flow if a matching trigger is found.

token: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
intent: book_appointment

📘

Entities

Matched entities will be available on flow scope using the entity name and also as a list by appending _list. For example the uttrance, "I love apples, pears and bananas." would result in the following flow scope data.

{
    "intent": "fruits"
    "fruit": "apples",
    "fruit_list": ["apples", "pears", "bananas"]
}

Data available in the flow scope

You can access details regarding the matched intent within within the flow scope.

luis

Matches an intent hosted on third party service luis.ai. Create an account and host your intents and Meya will make API calls.

Example

Create an app on luis.ai.

app_id: xxxxxxxxxxxxxxxxxxxxxxxxx
api_key: xxxxxxxxxxxxxxxxxxxxxxxx
intent: order_pizza

📘

Entities

Matched entities will be available on flow scope using the entity name and also as a list by appending _list. For example the uttrance, "I love apples, pears and bananas." would result in the following flow scope data.

{
    "intent": "fruits"
    "fruit": "apples",
    "fruit_list": ["apples", "pears", "bananas"]
}

Data available in the flow scope

You can access details regarding the matched intent within within the flow scope.

cms_nlu

If you've trained an NLU agent using Bot CMS, you can use the cms_nlu trigger to handle intents returned by whichever NLU provider is connected to the CMS space.

greetings

Greetings is a pre-trained NLP model (in English only) that can be used for your flows.

Intents

• hi
• how_are_you
• help
• whats_up

Example flow using the greetings trigger and auto routing.

route: true
states:
    hi:
        component: meya.random_text
        properties:
            responses:
                - "Hey there! 😀"
                - "Hi! 😃"
                - "Howdy! 😃"
                - "Hello! 😃"
        return: true
    how_are_you:
        component: meya.random_text
        properties:
            responses:
                - "I'm fine, thanks for asking 😃"
                - Great! How are you?
                - Super duper. You?
        return: true
    whats_up:
        component: meya.text
        properties:
            text: Not much. What's up with you?
        return: true
    help:
        component: meya.text
        properties:
            text: "Happy to help 😃"
        return: true

catchall

A catchall trigger will launch the flow whenever a user says something that hasn't matched to any of your other triggers. A flow triggered by the catchall trigger can be used to activate a fallback conversation flow.

Example - Detect Language

You can use the catchall trigger to detect a user's language. Start by creating a new flow language. Paste the following code into the flow.

states:
    first:
        component: meya.text
        properties:
            text: "Your language is {{ flow.language }}"

The above code results in the following behaviour. Note that in the first interaction detect_language is set to false, while in the following three interactions it was set to true.

📘

Tip: use a catchall trigger in your customer service bot and then transfer the request to a live agent using the human.transfer component as the first state of your flow.

start_chat

Triggers when someone opens the Meya Web chat for the first time. Checks to see if the user has opened the chat before by referencing the specified key. Optionally only fires on a URL specified in the settings.

Flow scope data

By default, the start_chat trigger populates some flow scope data, if available. If you want this data to persist beyond the current flow, use meya.set, or a Python component, to save the data to the user scope. Data on the flow scope expires once the flow is complete.

Here is a simple component that saves each of the above properties to the user scope. The second tab demonstrates how to use the component in a flow.

# -*- coding: utf-8 -*-
from meya import Component


class Save(Component):

    def start(self):
        for key in ["current_page_url", "user_agent", "referrer", "location", "browser_language"]:
            value = self.db.flow.get(key)
            self.db.user.set(key, value or "")
        return self.respond(message=None, action="next")

states:
    save_data:
        component: save_data

open_chat

Triggers anytime someone opens the Meya Web chat. Optionally only fires on a URL specified in the settings.

Flow scope data

By default, the open_chat trigger populates some flow scope data, if available. If you want this data to persist beyond the current flow, use meya.set, or a Python component, to save the data to the user scope. Data on the flow scope expires once the flow is complete.

Here is a simple component that saves each of the above properties to the user scope. The second tab demonstrates how to use the component in a flow.

# -*- coding: utf-8 -*-
from meya import Component


class Save(Component):

    def start(self):
        for key in ["current_page_url", "user_agent", "referrer", "location", "browser_language"]:
            value = self.db.flow.get(key)
            self.db.user.set(key, value or "")
        return self.respond(message=None, action="next")

states:
    save_data:
        component: save_data

👍

Tip

Use the open_chat event to keep track of what web page the user is on. Pro tip: Use custom query params to trigger different flows.

start_voice_chat

Triggers every time a voice chat start without a specific invocation.

📘

Integration support

Supported with Actions on Google

image

Triggers when someone enters an image. The image URL becomes available in flow.value

📘

Integration support

Supported on Messenger, Telegram, Twilio, Kik, Smooch and API & Webhook.

video

Triggers when someone enters a video.

📘

Integration support

Supported on Messenger, Telegram, Kik, Twilio and API & Webhook.

audio

Triggers when someone uploads an audio file.

📘

Integration support

Supported on Messenger, Twilio, Telegram and API & Webhook. Telegram also supports a voice recorded message.

file

Triggers when someone uploads a file.

📘

Integration support

Supported on Messenger, Twilio, Telegram and API & Webhook.

location

Triggers when someone enters their location 📍.

📘

Integration support

Supported on Messenger, Telegram, Smooch and API & Webhook.