SamTech 365

Power Platform, Copilot & SharePoint guides by Samir Daoudi

Create a knowledge base in QnA Maker

Step-by-step: create a knowledge base in Azure QnA Maker, then train, test and publish it so a chatbot can answer questions.
QnA Maker

In this article, I will take you through the steps you need to follow in order to create a Knowledge Base using QnAMaker.

Creating a KB in QnA Maker is the first step towards creating a bot or a virtual agent. You might have different reasons why you would want to have a virtual agent which responds to some predefined answers.

Whether it is to automate your recurrent requests in your help desk system or providing your users with an easy mean to interact with a bot and provide answers to frequently asked questions, it is always a good approach to have your users requests processed firstly by a bot, if this last is unable to provide the users with an acceptable response, your support team can jump in and provide the required support.

Create a knowledge base in QnAMaker.ai

  • Navigate to QnAMaker.io
  • Login using your Microsoft Account (you need to have an active Azure subscription to continue)
  • Once logged in, you will see the following home page

Create a knowledge base in QnA Maker

  • From the top navigation, select “Create a knowledge base“
  • This will take you through the following wizard

Create a knowledge base in QnA Maker – screenshot 2

  • The first step in this wizard, would be to create an Azure QnA Service (if you have not done this yet).
  • Click “Create a QnA Service“, which will open a new tab pointing to the Create a new resource in Azure, which is as follow

Create the QnA Service in Azure

Create a knowledge base in QnA Maker – screenshot 3

  • In this wizard you will need to provide the following details:
    • Subscription : Select in which subscription this resource will be created.
    • Resource Group : pick or create the resource group for the new resource.
    • Name : The name of the new resource.
    • Pricing tier : This will impact the performances, and the higher pricing tier you select, the better performances.
    • Azure Search Location : Geographical area where the search index will be stored.
    • Azure Search pricing tier : Same here, the higher the pricing tier, the higher indexing power you will have.
    • App name :  Provide a name to your App Service
    • Website location : The geographical location of your app service web site.
  • Provide and set your settings and click “Review + Create“.
  • Once your QnA Service created, go back to the QnA Maker tab and complete the wizard steps.

Pick your newly created QnA Service

  • In Step 2, select your newly created QnA Service

Create a knowledge base in QnA Maker – screenshot 4

  • In Step 3, just provide a name for your knowledge base

Create a knowledge base in QnA Maker – screenshot 5

Define your knowledge base source

  • Step 4 would be the most important, in which you will need to populate the following

Create a knowledge base in QnA Maker – screenshot 6

  • You can define the source of frequently asked questions, these can be brought from a remote url or from a structured file.
  • In this example, I am pointing to my other article in which I aggregated Teams frequently asked question.
  • Chit-Chat section defines how your bot will respond to the users and the behaviour of the bot.

Create a knowledge base in QnA Maker – screenshot 7

Well done ! you have created your first knowledge base.

Once done, you will see a screen with all your questions/answers, at this stage you have the possibility QnA Maker of amending these questions/answers or adding manually the ones you want.

Create a knowledge base in QnA Maker – screenshot 8

 

Test your knowledge base

From the main menu, you can click on “Test” button, which will open a chat window.

From this chat window, you can query your knowledge base and check the responses you get back.

Create a knowledge base in QnA Maker – screenshot 9

 

In the next article, I will describe how you can publish this knowledge base and integrate it in Teams !

Thanks for reading, don’t forget to like and comment😁

Related guides

Discover more from SamTech 365

Subscribe now to keep reading and get access to the full archive.

Continue reading