AI - Copilot: Basic & Advanced (Group 2)

​Agenda: Copilot 365, transforming productivity with AI

1. ​Introduction to the training path

   • ​Objective

   • ​Questions we are going to answer today

2. ​Introduction to the concept of AI

   • ​What is AI and what it can do

3. ​License: Copilot Free, 365 and Studio

   • ​What is a chatbot? and why do we use Copilot in this company?

   • ​The different types of Copilot

4. ​Author and Librarian

   • ​Grounding: the LLM + Microsoft Graph. How Copilot reads your files vs. its general training

   • ​The Author (LLM): Creative genius, access to resources on the internet

   • ​The Librarian (Microsoft Graph): Local expert, access to your environment

   • ​What you can do with Copilot in every Microsoft app

5. ​Referencing: The Laser Pointer (/)

   • ​The / (slash) command and the "Reference a file" button

   • ​How Copilot grounds its knowledge in specific documents to prevent generic answers

6. ​AI Loop

   • ​Why you can get stuck not using AI strategically, and how prompt engineering breaks the cycle

7. ​Prompt Engineering

   • ​What is a prompt and what elements matter

   • ​Prompting structure: objective, context, expectations

   • ​Interesting prompting techniques

​☕ 5-minute break

1. ​AI Prompting Framework

   • ​GSCE (Official Copilot Framework): Goal, Context, Source, Expectations

   • ​C.R.E.A.T.E (Expert Framework): Character, Request, Examples, Adjustments, Type of Output, Extras

   • ​Metaprompting: provide Copilot with your preferred framework and ask it to generate the prompt

2. ​Copilot Customisation

   • ​Personal Instructions: set your identity so you don't repeat your tone/role in every prompt

   • ​Chat Memory: build on previous answers within the same project

3. ​Practical Case

4. ​Agents

   • ​Knowledge base agents and when to create your own

   • ​Prebuilt Agents: Researcher and Analyst

   • ​Build your own agents

5. ​Closure

   • ​What have we learned?

   • ​Thanks and goodbye