Azure QnA Maker is a Microsoft Cognitive Service that primarily helps in providing a data sources as a knowledge base for a bot or application. QA Maker azure service uses cognitive services top of your data to build and train a simple data model and use them as question & answer model with natural languages. We have to first create the QnA Maker service in Azure before we can create the QnA Maker Knowledgebase. As Knowledgebase service required the reference of QnA Maker Azure service to start with.
Creating QnA Maker Azure Service
Sign in to Azure Portal and to create new Azure service, Select QnA Maker service from the list of Azure services.
You may choose this service from your saved list if it’s already added to “My Saved List” ? How – Check this out
In QnA Maker create blade provide all the required details and select the appropriate pricing tiers and regions for resources.
Click on Create, once all the provided fields are validated, it will start deployment of these resources in the selecting your subscription. The overall process will take a while.
Once the deployment is done, you will find the following resources created in your subscription.
Select the Cognitive Service type from the resource list and navigate to Quick Start option. You can start exploring the QnA Maker service and make use of the API call to create and manage the knowledge base.
Connecting QnA Maker Azure Service with Knowledge Base (KB)
To create a knowledge base, open QnAMaker.ai portal. Once login, you will find Create a Knowledge Base option in the top base. Select Create Knowledge base and move to the Step 2. Let’s skip the step 1 as it’s already completed.
When you move to, “Connect your QnA service to your KB” section, select the Azure Directory where we have to create the QnA Service, select the appropriate subscription and our newly created Azure QnA Service.
In the next step, provide the KB Name, and provide the details from where you need to populate the KB. In this case, we just use one simple Excel with few questions and answers.
Finally, complete the process by clicking on “Create your KB” button.
Wait for a while to crawl the content, extracting QnA pairs and creating the knowledge base. Once the knowledge base creation is done from the provided data source, you can test and train it, and finally, publish it.
Once you click on Publish, it will publish to ddntqnamaker azure services. Finally, click on Publish button to create the service endpoints.
This linked Knowledgebase with the QnA Maker Azure Service, and you can now connect the QnA Services with your BOT or any other application.
You can monitor the calls, errors, latency and data flow from the QnA Maker service dashboard.
Some Related Post :
- Rapid Development of Intelligent Bot Using Azure Bot Service
- Connecting Your Azure Bots with QnA Maker Services
- Setting Up Continuous Integration for Azure Bot Service
Hope this helps.