Introduction
In today’s digital landscape, FAQ bots play a crucial role in providing efficient customer support and disseminating information. In this article, we’ll explore how to create a custom FAQ bot without writing a single line of code. Our tools of choice include Azure Bot Service and Language Studio.
Development
1. Create the question-answering project
Sign in to the Azure Portal and create a Language Service resource.
Keep the default features and click on Continue to create your resource.
Choose the resource group, region, and pricing tier. Enter the resource name and check the box at the bottom. Then, click on Review + Create.
Once the resource is created, sign in to the Language Studio with your Azure credentials.
A popup dialog will appear. Choose the Azure directory, subscription, resource type (Language), and resource name (previously created). Click on Done.
Click on the Understand Questions and Conversational Language tab and select Custom Question Answering.
For this example, we don't have an Azure Search connected yet. So, click on Connect to Azure search. This will redirect to the Azure Search resources in Language Studio.
Click on Add Search Service.
Enter the service name and choose the location and pricing tier for the Azure Search service. Click on Create.
A confirmation dialog will appear. Click on Select Resource and delete projects.
Note. Pay attention to the message.
Once the resource is created, it will show this.
Return to the Language Studio tab. You may need to refresh this page for it to register the change to your resource. Select Create new project
Choose the option I want to set the language for all projects created in this resource, English as language, and click on Next.
Enter a project name, and leave the default answer with a setting of No answer found. Click on Next.
Review your choices and click on Create Project.
From the Manage sources page, select Add source, URLs.
Select Add URL and enter the URLs you want to add, then click on Add all: For this project, we used a URL with the news in Angular v18.
The extraction process takes a few moments to read the content and identify questions and answers. Questions and answers will determine if the underlying content is structured or unstructured.
After successfully adding the source, you can then edit the source contents to add more custom question-answer sets.
2. Test your project
Select the link to your source, this will open the edit project page.
Select Test from the menu bar. Enter a question. An answer will be generated based on the question-answer pairs that were automatically identified and extracted from your source URL:
If you check the box for include short answer response, you will also see a precise answer, if available, along with the answer passage in the test pane when you ask a question.
Select Inspect to examine the response in more detail. The test window is used to test your changes to your project before deploying your project.
From the Inspect interface, you can see the level of confidence that this response will answer the question and directly edit a given question and answer response pair.
3. Deploy your project
From the sidebar, select Deploy knowledge base to enter the Deploy project menu.
When you deploy a project, the contents of your project move from the test index to a prod index in Azure Search.
Select Deploy, and your project is now successfully deployed. You can use the endpoint to answer questions in your own custom application or in a bot.
4. Create a bot
After deploying your project, you can create a bot from the Deploy project page. Click on Create a bot.
A new browser tab opens for the Azure portal, with the Azure AI Bot Service's creation page. Configure the Azure AI Bot Service in the Basics tab. Click on Next.
In the Web App tab, you will need the key for the Language Service. Go to the Language Service you created, and in the Keys and Endpoint section, copy one of the two keys.
Once you pasted the key, click on Review + Create.
Click on Create.
After the bot is created, go to the Bot service resource. Under Settings, select Test in Web Chat.
Select Channels in the Bot service resource that you have created. You can activate the Bot in additional supported channels.
Conclusion
We’ve demonstrated how to build a powerful FAQ bot without writing code. By seamlessly integrating Azure Bot Service, Language Studio, and content from us, you can create a personalized and efficient solution for your users. Encourage readers to explore further and customize their own no-code FAQ bots.
Thanks for reading
Thank you very much for reading. I hope you found this article interesting and may be useful in the future. If you have any questions or ideas you need to discuss, it will be a pleasure to collaborate and exchange knowledge.
Join the conversation! Your thoughts help the community grow.