Microsoft 365 and Copilot

Jul 1, 2024

Lecture Notes: Microsoft 365 and Copilot

Key Topics

  • Microsoft 365 commitment to data security and privacy
  • Overview of Copilot
  • Large Language Models (LLMs)
  • Integration with Microsoft Graph
  • Data Privacy and Security

Microsoft 365 Commitment

  • Built on data security and privacy commitments
  • Compliance with GDPR and EU Data Boundary

Copilot Overview

  • AI-powered productivity engine
  • Uses LLMs with data in Microsoft Graph
  • Embedded in Microsoft 365 apps like Word, PowerPoint, Teams
  • Provides real-time intelligent assistance

Large Language Models (LLMs)

  • AI algorithms using deep learning and vast datasets
  • Understand, summarize, predict, and generate content
  • Pre-trained models (e.g., GPT-4 Turbo)
  • Learn from data sources like books, articles, web pages, images
  • Generates text based on preexisting patterns

Copilot and LLMs

  • Uses an approach called grounding to process queries
  • Accesses content/context via Microsoft Graph
  • Combines user documents, emails, calendar, chats, meetings, contacts
  • Generates accurate, relevant, contextual responses

Azure OpenAI Services

  • Azure OpenAI services used, not public OpenAI
  • No data collection or storage by Azure OpenAI
  • No caching of customer content or modified prompts

Summary

  • AI assistant for work
  • Coordinates LLMs, Microsoft Graph content, Microsoft 365 apps, and web
  • Provides real-time intelligent assistance
  • Enhances creativity, productivity, and skills
  • Adheres to privacy and security standards

Learn More