LUIS Intent Vectors provide a robust and configurable connection between your trained LUIS models and Flow Skills. Intent Vectors simplify the work it takes to develop the dialog required to fulfill an intent by mapping LUIS entities to configured entities in AtBot. This allows your bots to automate any required dialog with the user to capture missing entities when an intent is used.
Creating Intent Vectors
On the LUIS Intent Vector screen click Create Intent Vector, fill out the fields below and click Create Intent Vector
LUIS App Region
Select the region your LUIS app is deployed to
LUIS API Key
Enter your LUIS Author key
Select the app that contains the intent you wish to configure
Select the intent you wish to configure
Intent Vector Settings
The following are settings displayed when an Intent Vector is selected.
The read-only name of the App from LUIS
The read-only name of the Intent from the LUIS App
LUIS API Key
The programmatic API key for authoring your LUIS apps. This key is used in the AtBot admin to load in your models.
LUIS Endpoint Key
The programmatic API Key for executing LUIS calls. This key is used by the bot to make calls to your LUIS endpoint. This can be a paid key.
The message that will be sent to the user if the intent to cancel is found in the middle of an Intent Vector process
Optionally add additional Contributors to the Intent Vector. Contributors will be able to edit all settings of the IV as well as select it in Flow. Search for contributors by UPN.
For each entity that you'd like captured with your intent, add that entity from the available entities and configure how its value will be captured in dialog.
The editable name for the property that will be available in Flow
The data type of the entity. This will also determine how the data will be collected from the user (Person, Date, Choice)
The read-only name of the entity in LUIS
The description of the property. This will show in the properties window in Flow
The question to be asked of the user to collect this piece of data
If checked, this entity will be required to be collected from the user. If it isn't in the original utterance, the bot will ask using the Follow-up Question
If checked, the entity will be put into a collection and allow for multiples to be mentioned in the utterance