LUIS - Language Understanding Intelligent Service - step by step instruction

LUIS

Here is the definition of LUIS for Microsoft;

"

LUIS is a machine learning-based service to build natural language into apps, bots, and IoT devices. Quickly create enterprise-ready, custom models that continuously improve.

"

We can use LUIS in our chatbot solution in order to convert human-readable sentences of customer to a machine-readable format. LUIS generates JSON format through utterences from customer free-text input.

The ability to understand what your user means conversationally and contextually can be a difficult task, but can provide your bot a more natural conversation feel. Language Understanding, called LUIS, enables you to do just that so that your bot can recognize the intent of user messages, allow for more natural language from your user, and better direct the conversation flow. This topic walks you through setting up a simple bot that uses LUIS to recognize a few different intents.

Here are the step by step instructions of LUIS configuration and implementation. 

1.     Sign in to LUIS through  https://www.luis.ai to create your own app.
You can create and manage your applications on My Apps section from the interface.

Note: you should to have MSDN or Microsoft Azure subscriptions in order to login the portal.

2.      You should grant access to your information for LUIS by clicking “Yes” (Evet).

3.      Select region then accept agreement accordingly then click Continue for the next step.

4.      You will be forwarded to the main page.

5.      Click “Create a LUIS app now”

6.      “My Apps” section will display

7.      Select “Import new app” from the menu.

8.      Click “Choose app file (JSON format) …” from the popup window

9.      Select reminders.json file located in the CognitiveModels folder of the sample.
In the Optional Name, enter LuisBot. This file contains three intents: Calendar_Add, Calendar_Find, and None. We'll use these intents to understand what the user mean when they send a message to the bot.

Notice : You should download botbuilders-tools from https://github.com/Microsoft/botbuilder-tools/tree/master/packages/LUIS for latest version or open attached zip file.

botbuilder-samples (It was taken 30th December,2018)

10.      Train the app. 
how to : https://docs.microsoft.com/en-us/azure/cognitive-services/LUIS/luis-how-to-train). 

      You can add your intents as described following URL : https://docs.microsoft.com/en-us/azure/cognitive-services/LUIS/luis-how-to-add-example-utterances

11.      Imported sample has defined utterances. It is possible to add new capability by teaching new utterances through created LUIS application.

12.      Give a name. For example; Subscriptions then click “Done”.

13.      Subscriptions predication under app. will be displayed.

14.      Type possible utterances then click enter.

15.      Click train from the menu bar

16.      Once train is completed, button will turn into green light.

17.      Click Publish to make it ready.

18.      Click Keys and Endpoints in order to see necessary integration endpoints.Tip: you can activate Bing for spell check

Summary : 

By using LUIS of the Microsoft Cognitive Service, you will be able to make smarter system. It can be used by Call Center or Customer Relationship Management (CRM) applications.

Handling customer request through an interface (like chatbot) which is bound to LUIS, will increase customer satisfaction without occupying agents that increases effectiveness and cost cutting .

enjoy...