Links
Last updated
Was this helpful?
DialoX Manual
Last updated
Was this helpful?
Nodes can be connected using Links. When you draw a line from one node to another it will create a link between the two. There are two kinds of special links. They are both only available in Ask node types:
Labelled links
Dialog triggers
The available choices are managed by drawing lines (= node links) from this node to other nodes. When the line is drawn without a target node, the system will automatically ask to choose a node type to create for you. The link type will be automatically set to "labelled".
A "labelled" link represents the expected choices that will satisfy the question. Every link represents one choice.
Labelled links are mainly used for the node Ask: Choice. A node link consists of the following details:
Label
The label of the choice.
Will be used to match the input of the user to the right choice. On visual channels the choices are represented as quick replies. On voice channels the choices are also available as DTMF choices.
Strict matching
The label is matched strictly.
So when the label is "A" the choice will not be triggered when "AA" is given if Strict matching is enabled.
Intent
To be able to match the choice to the user input using a trained user intent.
DTMF Key
To be able to use a DTMF key on your phone to select the trigger.
DTMF are the keys on your phone, therefore this option will only be relevant for bots that answer the phone.
When Strict Matching is enabled the platform will be strict when using the label for matching the trigger. This means, when for instance, the label is "Sales", the trigger will not match if the user says "Sales Departments". It will only match when the user says specifically and only the word "Sales".
This is usefull for situations where you have two or more choices where one of them is partly used in the others. So for instance when there are two choices: "Sales" and "Sales Questions". Then the system will always match "Sales", even when the user says "Sales Questions".
Dialog triggers can be used on Ask nodes to handle user input or events that do not belong to the main choices of the Ask, the Labelled links. Dialog triggers can trigger by user intents or by system events.
The following example shows two types of dialog triggers. The first is using a trained user intent and the second an unknown event.
The user intent for "Why?" can be used like this:
The trigger links have slightly different properties that can be configured. It shares the intent property, but instead of a label it offers a system dialog (to handle the "unknown" situation)
Intent
To be able to match the choice to the user input using a trained user intent.
System dialog
To be able to respond to system dialogs, like unknown
Unknown is used to handle any unmatched user input
Once the flow that is triggered by the dialog trigger is finished, it will return to the original Ask it came from.
Dialog triggers that trigger on an intent are like "hidden" choices. These choices will not satisfy the Ask. Instead they will execute the nodes that are linked and then directly return to the Ask, prompting the Try again question. This function can be very helpful for explanatory dialogs, like the above "Why?".
Certain system events can also trigger another dialog. These are:
Unknown
Max unknowns
Timeout
The trigger event Unknown can be very helpful to handle unmatched user intent so proper fallback can be arranged like explaining what is needed or escalating to a user.
Max unknowns is the event that triggers after a 3 times an unknown triggers has fired. This is to prevent users from endless looping in an Ask node.
Another way for users to get stuck in an Ask is by not choosing at all. Sometimes you want the bot to continue anyway when this happens. For this purpose the Dialog trigger "Timeout" can be used. On WhatsApp the timeout is 5 minutes. On phone it is relative to the input expected, but count on around 8 to 15 seconds. On other channels it is 2 minutes.
User intents can be trained in the of the platform.
User intents can be trained in the section of the platform.
