1. Understanding Microsoft Copilot
Welcome to this course on Microsoft Copilot! I'm Nia, an AI Product Builder at Microsoft, and I'll be your host. Let's dive in!
2. Course overview
While this course focuses on Microsoft 365 Copilot, we'll briefly cover Copilot - a web-based digital assistant - in the next video.
3. Course overview
We'll then explore Copilot's integration with Microsoft 365 applications to enhance productivity
4. Course overview
and conclude with Responsible AI and best practices.
The images and videos in this course were created on a Windows machine; some images are AI-generated.
5. Context switching
Managing different projects can mean switching between many applications, emails, messages, and documents. It can be hard to focus, and it can feel overwhelming.
6. Copilot as an assistant
Enter Copilot, our project sidekick. Copilot can review emails and documents, summarize email threads, highlight key points, and retrieve documents across applications.
7. Copilot as an assistant
With Copilot handling the busywork, we can focus on tasks that require creativity and decision-making.
8. Microsoft Copilot vs. Microsoft 365 Copilot
There are two types of Copilot, each serving different purposes. The first is a web-based digital assistant that performs web searches and generative tasks. The second is integrated into Microsoft 365 applications like Word, Excel, and Outlook, and helps with app-specific work.
Throughout this chapter, we'll refer to the first version as Microsoft Copilot and the second as Microsoft 365 Copilot.
9. Microsoft Copilot
Microsoft Copilot can assist with tasks like web searches, creating images, and generating content, such as writing stories.
10. Microsoft 365 Copilot
Microsoft 365 Copilot is designed to enhance productivity within Microsoft applications, powered by the broader Copilot system and enterprise-grade data security.
11. Copilot system
The Copilot System is Microsoft's orchestration engine. It's made of three parts: Microsoft 365 apps, Microsoft Graph, and Large Language Models or LLMs.
12. Copilot system - Microsoft 365 apps
The first components are Microsoft 365 apps like Word, Excel, and Outlook. To interact with Copilot, we first type our request, which is the prompt.
13. Copilot system - Microsoft Graph
After we make the request, Microsoft Graph steps in. Think of this as a map that connects all our data like emails, shared documents, calendars, and more. This gives Copilot the background it needs to understand our request. For example, if we ask Copilot for a summary, Microsoft Graph helps it find related information across applications.
14. Copilot system - Large Language Models
Once Copilot has the background from Microsoft Graph, it sends the request to the Large Language Model, which is trained on publicly available data to understand and generate human-like text. The LLM uses what it learned from Graph, and its language skills, to write, edit, or answer questions.
15. Copilot system
This is an iterative process: after the response is generated, Microsoft Graph checks that the information returned is relevant.
16. Copilot system
The response is then sent back to the 365 applications, and can be read by the user.
17. Copilot system
In short: we interact with Copilot, the Graph provides context, and the LLM generates a response.
18. Let's practice!
Let's practice!