Create your first language model

Language models are fundamental components for configuring your Health Bot experience. Language models interpret end user utterances and trigger the relevant scenario logic in response.

In this tutorial we will follow the steps required to create an NLP model using LUIS. Our model will understand when a customer is asking to schedule an appointment and trigger a scheduling scenario.

Create your LUIS model

  1. You should navigate to your management portal and create a new application. In the portal create a model that understands when an end user is asking to schedule an appointment.
  2. Define the intent as “Schedule appointment” and add several example utterances for the model. Learn more about how to create a LUIS application

Screen shot of the LUIS application portal and test utterances

  1. You will need the Application ID for your new model. In the LUIS portal navigate to Application Information >> Application ID

Screen shot of LUIS application ID

  1. You will also need the Authoring key (also known as a Subscription Key). Navigate to Key and endpoint settings >> Authoring key

Screen shot of LUIS subscription key

Create the Health Bot Service Language model

  1. Login to the Health Bot Management portal and navigate to Language >> Models. Here you should add a new model and select the LUIS recognition method.

Screen shot of language model recognition methods

  1. Provide the required name and description fields. These can be any name you like and are only seen internally to recognize the model.

  2. Region: This should match the deployment region you have selected for your LUIS application.

  3. Provide the Application ID and Authoring key from the LUIS application as per steps three and four.

Screen shot of LUIS application credentials

  1. The Bing, Staging and Verbose fields are additional features that you can enable for your LUIS application. For this tutorial you can leave the defaults.

  2. If the credentials have been provided correctly you will be able to fetch the intents from your LUIS application. You should then map the intent to your custom scheduling scenario.

Screen shot of LUIS application credentials

You can now create the language model and test it. Click save and open the webchat in the management portal. Type “I would like to schedule an appointment”. You should see that your custom scenario is being triggered.

Screen show of webchat and testing a language model