How Copilot for Microsoft 365 Works

  • Thread starter Thread starter Mark Kashman
  • Start date Start date
M

Mark Kashman

Knowing how technology works eases your mind, minimizes confusion, and deepens your understanding as you incorporate it into your daily work routine.



Copilot for Microsoft 365 leverages large language models (LLMs) that interact with your data using the Microsoft Graph and optionally the Internet to generate personalized experiences with related context as you work. Copilot for Microsoft 365 provides a powerful orchestration engine and combines the power of Large Language Models, your universe of data via Microsoft Graph, the Internet, and your Microsoft 365 apps. Importantly, Copilot for Microsoft 365 only uses information that you explicitly have access to. It only remembers your conversation while it’s in that conversation and we don’t use your information to train the Large Language model. The chat history is wiped clean with each new conversation.



OK, let’s go behind the scenes to further understand the architecture and dataflows behind Copilot for Microsoft 365, including retrieval augmented generation techniques, app commanding, and options to securely connect to external data. You’ll see the potential within chat and how it can save you time with examples of specific use cases.



Watch | “How Copilot for Microsoft 365 Works” presented by A.J. Brush and Mary David Pasch at the recent Microsoft 365 Community Conference #M365Con:






large?v=v2&px=999.pngPowerPoint slide screenshot "Copilot for Microsoft 365 architecture" as presented at the Microsoft 365 Community Conference | April 30 – May 2024 | Orlando, Florida.

More on demand | Recent long-form Copilot and AI videos from Microsoft




Additional, important Copilot and AI resources from Microsoft




Thanks for your time discovering more about Copilot for Microsoft 365, Karuana Gatimu and Mark Kashman

Continue reading...
 
Back
Top