DEV Community

Mathieu Kessler
Mathieu Kessler

Posted on

365 Microsoft Copilot Prompts I Wish I Had When I Started

The $30/month disappointment

Six months ago, my company rolled out Microsoft 365 Copilot. $30/user/month. The pitch deck looked amazing: "AI-powered productivity!" "Work smarter!" "Automate everything!"

Week 1 reality check:

Me: "Help me with my emails"
Copilot: "I'd be happy to help! What would you like to know about emails?"
Me: 😐
Enter fullscreen mode Exit fullscreen mode

The generic prompts from blog posts didn't work. ChatGPT tutorials were useless. I was paying $30/month for a fancy chatbot that couldn't access my actual work.

The problem wasn't Copilot. It was the prompts.

Why your ChatGPT prompts fail in Copilot

Here's what nobody tells you about M365 Copilot:

ChatGPT doesn't know about your work. Copilot does.

Feature ChatGPT M365 Copilot
Your emails ❌ Can't access ✅ Reads Outlook
Your files ❌ Copy-paste only ✅ Accesses SharePoint
Your data ❌ None ✅ Excel, Teams, OneDrive
Actions ❌ Text only ✅ Creates docs, sends emails

This means ChatGPT prompts like "write me an email" are useless in Copilot.

But M365-optimized prompts? Game-changer.

The difference between generic and enterprise prompts

Generic prompt (doesn't work):

"Summarize my emails"
Enter fullscreen mode Exit fullscreen mode

Result: Copilot has no idea which emails, what context, or what format you need.

Enterprise-optimized prompt:

"Analyze emails from the last 7 days about Project Phoenix. 
Create a summary table with:
- Key decisions made
- Action items (with owners)
- Budget impacts
- Blockers

Format for executive audience (my VP reads this)."
Enter fullscreen mode Exit fullscreen mode

Result: Copilot accesses Outlook, analyzes 47 emails, creates formatted table, identifies 12 action items, flags 3 budget concerns. Takes 30 seconds instead of 2 hours.

That's the difference.

5 prompts that changed how I work

After 6 months of daily Copilot use, these are my MVPs:

1. The Email Intelligence Prompt (Outlook)

Problem: Drowning in email threads with 15+ people.

Prompt:

"Review the email thread with subject '[Subject Line]'. 
Extract:

DECISIONS MADE:
- [List with dates and who decided]

ACTION ITEMS:
- [Task] - Owner: [Name] - Due: [Date]

DEPENDENCIES:
- [What's blocking progress]

CONTEXT I MIGHT HAVE MISSED:
- [Important details from earlier in thread]

Format as a scannable brief (I need to respond in 5 minutes)."
Enter fullscreen mode Exit fullscreen mode

Time saved: 15-20 minutes per complex thread.

2. The Meeting Prep Assassin (Teams + Outlook)

Problem: Back-to-back meetings, no time to prep.

Prompt:

"I have a meeting about [Topic] in 10 minutes with [Names].

Search my emails and Teams chats from the last 2 weeks. 
Create a brief with:

WHAT THEY CARE ABOUT:
- [Their priorities based on communications]

OPEN QUESTIONS FROM LAST TIME:
- [Unresolved items]

DECISIONS NEEDED TODAY:
- [What we need to decide]

LANDMINES TO AVOID:
- [Sensitive topics based on email tone]

Keep it under 200 words. I'm reading this in the elevator."
Enter fullscreen mode Exit fullscreen mode

Time saved: Would otherwise go in blind. This takes 20 seconds.

3. The Data Detective (Excel)

Problem: Messy Excel file, need insights fast.

Prompt:

"Analyze [Sheet Name] in this Excel file.

Find:
1. TOP 3 ANOMALIES (things that look wrong/unusual)
2. HIDDEN PATTERNS (correlations I might miss)
3. RECOMMENDED ACTIONS (what should I investigate)

Then create a summary table showing:
- Total by [Category]
- Growth rate vs last [Period]
- Top 5 and Bottom 5 performers

Explain like I'm not a data analyst (because I'm not)."
Enter fullscreen mode Exit fullscreen mode

Time saved: 45 minutes of manual analysis.

4. The Document Synthesis Machine (Word + SharePoint)

Problem: Need to compare 5 different policy documents.

Prompt:

"I need to compare these documents: [List files in SharePoint]

Create a comparison table showing:
- WHAT CHANGED between versions
- CONFLICTS (where they contradict each other)  
- GAPS (what's covered in one but not others)
- RECOMMENDED APPROACH (which version to use for what)

Highlight any legal/compliance differences in RED.

I'm presenting this to leadership in 30 minutes."
Enter fullscreen mode Exit fullscreen mode

Time saved: This used to take me 3 hours. Now takes 2 minutes.

5. The Cross-App Workflow (Excel → PowerPoint → Outlook)

Problem: Create exec dashboard from data, present it, send follow-ups.

Prompt:

STEP 1 (Excel):
"Create a pivot table from [Sheet] showing revenue by region and product. 
Add a trend chart for Q3-Q4. Highlight top 3 and bottom 3 performers."

STEP 2 (PowerPoint):
"Create a 3-slide exec summary:
Slide 1: Overall performance (use the Excel chart)
Slide 2: Top 3 wins
Slide 3: Bottom 3 concerns + recommended actions

Use our corporate template. Keep it punchy."

STEP 3 (Outlook):
"Draft an email to leadership with:
- Subject: Q4 Revenue Analysis - Key Insights
- Body: 3-sentence summary + attached deck
- Tone: Confident but flag concerns
- Ask: 'Can we discuss bottom 3 in Tuesday's call?'"
Enter fullscreen mode Exit fullscreen mode

Time saved: 4-6 hours of work automated in one flow.

How I organized 365 prompts

After building these workflows for 6 months, I realized: nobody has a comprehensive M365 Copilot prompt library for enterprise work.

So I built one.

The structure:

🟢 Quick Start (50 prompts)

  • Work in any Copilot tier (even free)
  • No M365 data access needed
  • Universal workflows (writing, learning, planning)

🟡 Advanced (315 prompts)

  • Require M365 Copilot or Copilot for Enterprise
  • Organized by:
    • App: Outlook (70), Excel (40), PowerPoint (35), Word (40)
    • Role: Sales, Engineering, HR, Finance, Legal, etc. (12 roles)
    • Use case: Email intelligence, data analysis, document synthesis

Key features:

GCSE Framework examples (Microsoft's official prompting methodology)

Cross-app workflows (the magic nobody talks about)

Troubleshooting guide (when Copilot refuses or hallucinates)

Before/after comparisons (so you see the difference)

What makes it different:

Most prompt libraries are:

  • Generic ("write me an email" 🙄)
  • ChatGPT-focused (not M365-specific)
  • Beginner-level only

This library is:

  • Enterprise-grade (actually used in production)
  • M365-optimized (leverages Outlook, Excel, Teams integration)
  • Multi-level (beginners → power users)

The GitHub repo

I open-sourced everything: 365 prompts, fully categorized, ready to use.

🔗 kesslernity/awesome-microsoft-copilot-prompts

What you get:

📁 awesome-microsoft-copilot-prompts/
├── 📄 README.md (navigation + GCSE framework)
├── 📁 prompts/
│   ├── 🟢 QUICK-START.md (50 universal prompts)
│   ├── 📧 outlook.md (70 email prompts)
│   ├── 📊 excel.md (40 data prompts)
│   ├── 📊 powerpoint.md (35 presentation prompts)
│   ├── 📝 word.md (40 document prompts)
│   ├── 💼 sales.md (role-specific)
│   ├── ⚙️ engineering.md (role-specific)
│   └── ... (12 role categories)
Enter fullscreen mode Exit fullscreen mode

How to use it:

  1. Browse by role (Sales, Engineering, HR, etc.)
  2. Browse by app (Outlook, Excel, PowerPoint, Word)
  3. Browse by use case (Email intelligence, data analysis)
  4. Copy prompt → Paste in Copilot → Customize for your context

No signup. No paywall. Just prompts that work.

What I learned building this

1. Specificity beats cleverness

Bad prompt: "Help me with this email"

Good prompt: "Analyze this email thread. Extract decisions, action items, blockers. Format for my VP."

2. Context is everything

Copilot works best when you tell it:

  • GOAL: What you're trying to achieve
  • CONTEXT: Who the audience is
  • SOURCE: Which files/emails to use
  • EXPECTATIONS: What format you need

(This is Microsoft's GCSE framework - included in the repo)

3. Cross-app workflows are magic

The real power isn't in individual prompts.

It's in chaining them: Excel → PowerPoint → Outlook in one flow.

Most people never discover this because tutorials focus on single-app usage.

4. Enterprise ≠ Consumer

Consumer AI: "Make this sound professional"

Enterprise AI: "Analyze 47 emails about budget overruns. Flag compliance risks. Create action plan by department."

Totally different use cases.

The prompts I'm still building

I'm actively maintaining this repo. Here's what I'm working on:

🚧 Coming soon:

  • Compliance & audit workflows
  • Multi-language support (international teams)
  • Industry-specific prompts (healthcare, finance, legal)
  • Integration with Power Automate
  • Video tutorials for complex workflows

What am I missing?

I built this based on my experience, but I'm one person. What prompts would help YOU?

💬 Tell me in the comments:

  • What's your role?
  • What's your biggest Copilot pain point?
  • What workflows take you the most time?

I'll add the most-requested prompts to the repo.

Try it yourself

If you use Microsoft 365 Copilot (or are about to start):

  1. ⭐ Star the repo if you find it useful (helps others discover it)
  2. 📖 Browse the prompts that match your role
  3. 💬 Open an issue if you have a prompt request
  4. 🔀 Submit a PR if you have prompts to contribute

The goal: Make M365 Copilot actually worth the $30/month.

🔗 GitHub: awesome-microsoft-copilot-prompts


Why I built this

Six months ago, I was frustrated with Copilot.

Generic prompts didn't work. ChatGPT tutorials were useless. I felt like I was paying $30/month for nothing.

Then I learned how to write enterprise prompts.

My email time went from 2 hours/day → 30 minutes/day.

My meeting prep went from 20 minutes → 20 seconds.

My status reports went from 1 hour → 5 minutes.

I got 10+ hours back every week.

But I had to figure this out the hard way. Trial and error. Lots of error.

So I documented everything. 365 prompts. All the workflows that work. All the mistakes to avoid.

Now you don't have to learn the hard way.

Just use the prompts. Save 10 hours/week. Actually get value from your Copilot license.

And if you find it useful? Star the repo. It helps other people discover it.

🔗 kesslernity/awesome-microsoft-copilot-prompts


P.S. What prompts should I add next? Drop a comment below. 👇


Follow me for more enterprise AI content:

Top comments (0)