📎 Webclip
DIY FAQ Bot
The post walks through a basic chatbot setup using Azure Bot Service, QnA Maker, and Twilio. It frames the bot as a way to handle noisy inbound messages by answering common questions from a FAQ-style knowledge base.
Reading notes#
- The post describes a simple chatbot built with minimal programming.
- It says the bot was meant to help manage DMs and common inbound messages.
- The author groups messages into categories such as Azure questions, career advice, speaking inquiries, close-friend conversations, creepy greetings, and miscellaneous opportunities.
- It says QnA Maker can create a bot from a simple FAQ list.
- It instructs the reader to create a knowledge base in the QnA Portal.
- It says the knowledge base is a collection of searchable questions and answers hosted in Azure.
- It notes that service details include name, subscription, location, pricing tier, resource group, and optional App Insights.
- It says the knowledge base is populated by uploading a Word document of FAQs.
- It says the bot personality can be set with a chit-chat option such as friendly.
- It says the bot can be tested before publishing.
- It says publishing turns the knowledge base into an endpoint.
- It says an Azure Web App Bot is created with SDK v3, C#, and the Question and Answer template.
- It says the bot is connected to the published QnA service through application settings.
- It says Twilio is used as the channel for SMS messaging.
- It says the Twilio setup needs a phone number, Account SID, and Auth Token.
- It says a TwiML app is created with the messaging request URL https://sms.botframework.com/api/sms.
- It says the Twilio credentials are added in Azure Channels.
- It says the result is a basic QnA bot that can answer common questions over text.
