COPILOT PROMPT LIBRARY - STARTER VERSION (to personalize)
Recommended structure: Role + Task + Context + Format (RTCF)

=== DRAFTING (Word) ===
1. Turn these notes into a one-page [document type]: title, context, decision, terms, contact. Professional and direct tone.
2. As a [your role], write [document] for [audience] to [goal]. Format: [length, structure]. Tone: [formal/warm/direct].
3. Rewrite this passage half as long, without losing information.
4. Rewrite for a non-specialist reader, short sentences, zero jargon.

=== SUMMARIZING ===
5. Summarize this document in 5 key points for a busy executive.
6. Make a table of the actions mentioned: action, owner, due date.
7. What does this document say about [topic]? Cite the passages.
8. Compare /[file A] and /[file B]: list the differences and contradictions.

=== EXCEL ===
9. Analyze this data: the 3 main findings, then any outliers.
10. Add a [name] column: [logic in plain English, with the exact column names].
11. Explain this formula step by step: [paste the formula].
12. Create a chart of [measure] by [dimension]; iterate: type, filters, labels; final title: [your message].

=== EMAILS (Outlook) ===
13. Write a reply that thanks, answers the questions asked and sets a due date of [date]. 5 lines max, tone [x].
14. Politely decline [request], propose [alternative], warm but firm tone.
15. [Coaching] Assess the tone of this draft: how will it be perceived? Risks of misunderstanding? Suggestions.

=== MEETINGS (Teams) ===
16. Summarize what I missed in 5 points.
17. List the decisions made and the actions with owner and due date.
18. Prepare me for my [time] meeting: context, participants, open points, documents to re-read.

=== CROSS-CUTTING (Work mode) ===
19. Summarize what happened on [project] over the last two weeks, with sources.
20. Which recent documents mention [topic] and who wrote them?

REMINDERS
- Check the clickable references; never send raw AI output.
- Company data: only in Copilot 365 (work account).
- Iterate rather than retype; give an example of the expected result.