Microsoft Bot Framework Cheat Sheet

• Powered by Azure AI Bot Service
• A comprehensive framework for building intelligent, enterprise-grade conversational AI experiences across multiple channels.
• Enables natural language understanding, multi-turn dialogues, and proactive messaging for a seamless conversational experience.
• Integrates natively with Azure AI services like Azure AI Language, Azure AI Speech, and Azure OpenAI for enhanced bot intelligence.
• Supports deployment across channels including Microsoft Teams, Web Chat, Facebook Messenger, Slack, and custom clients via Direct Line.

Important Terminologies

• Bot
  A conversational application that interacts with users through text, cards, or speech to perform tasks or provide information..
• Activity
  The basic unit of communication between the bot and the user. It can either be a message, a command, or a conversation update.
• Turn
  A single interaction that includes one message from the user and one response from the bot.
• Dialog
  A structured component used to manage conversation flow. It supports multi-step, multi-turn interactions with branching logic.
• Channel
  A platform or medium (e.g., Microsoft Teams, Facebook Messenger, Direct Line) where users engage with the bot.

Key Features

1. Bot Development Tools

• Bot Framework SDK: Develop bots using open-source SDKs for .NET, JavaScript, and Python. Includes support for dialogs, adaptive cards, and rich message handling.
• Bot Framework Composer: Visual authoring canvas to design and manage dialogs, triggers, and language generation (LG) templates without heavy coding.
• Bot Framework CLI: Command-line tools to automate bot management tasks like publishing, QnA setup, and configuration.

2. Natural Language Understanding (NLU)

• Azure AI Language: Integrate state-of-the-art intent and entity recognition for natural language processing in bots.
• Conversational Language Understanding (CLU): A feature of Azure AI Language that enables bots to understand user intent and extract key information from natural language input
• Orchestrator: Route user input across multiple LUIS apps or QnA knowledge bases using transformer-based models.

3. Conversation Management

• Dialogs: Manage multi-turn conversations using waterfall dialogs, adaptive dialogs, and declarative dialog definitions.
• Memory Scopes: Store and manage user, conversation, and turn-specific data within dialog contexts.
• Interruptions & Event Handling: Handle unexpected inputs, cancellations, or proactive messages gracefully during conversations.

4. Proactive Messaging

• Bots can initiate conversations based on external triggers (e.g., alerts or scheduled events).
• Supports storing conversation references and sending messages outside of user-initiated sessions.

5. QnA and Knowledge Integration

• Azure AI Language Question Answering: Build knowledge-based bots that respond to FAQs using unstructured or semi-structured content.
• Combine QnA with NLU for hybrid bots that can switch between knowledge and task-based interactions.

Integration with other Azure services

Sample Use Cases

• Handle FAQs, troubleshoot issues, escalate to human agents, and assist with internal IT, HR, or administrative support.
• Help users schedule appointments, manage tasks, navigate applications, and access personalized information.
• Guide users through product discovery, provide recommendations, handle purchases, and track orders.
• Support sensitive or regulated interactions such as checking balances, scheduling medical appointments, delivering health or government service information, and sending alerts.
• Deliver course materials, manage enrollment or scheduling, answer academic questions, and guide users through learning paths.
• Survey, Feedback, and Notification Collect user feedback, conduct polls, and send proactive updates such as reminders, alerts, or announcements.

Pricing

• Azure AI Bot Service offers the following:
  • Free tier for standard channels.
  • Premium channels like Direct Line and SMS are billed per message or session.
• Direct Line is priced based on the number of messages exchanged between the user and the bot.
• Azure AI Language (CLU) charges based on the number of text records analyzed for intent and entity recognition.
• Azure OpenAI Service is billed per token (input and output) used during text generation.
• Azure AI Speech is priced based on the duration of audio processed for speech-to-text or text-to-speech conversion.

Microsoft Bot Framework References:

https://learn.microsoft.com/en-us/azure/bot-service/bot-service-overview
https://learn.microsoft.com/en-us/azure/bot-service/bot-builder-basics
https://learn.microsoft.com/en-us/azure/bot-service/bot-service-quickstart-create-bot
https://azure.microsoft.com/en-us/pricing/details/bot-services/