Skip to content

📊 AI-in-One Dashboard — Complete Setup & User Guide

Field Detail
Priority 🔥 Urgent
Status ✅ Deployed
Category 💼 Official / 🎓 Learning Project
Date added 2026-03-29
Completed 2026-03-29
Source microsoft/AI-in-One-Dashboard
Local repo C:\ssClawy\AI-in-One-Dashboard
Lab tenant M365CPI52224224.onmicrosoft.com

What Is the AI-in-One Dashboard?

☕ Simple analogy: Imagine you own 50 café branches and you've just installed an AI barista (M365 Copilot) in each one. You need to know:

  • How many customers are using the AI barista? → Adoption
  • Which branches use it the most? → Engagement
  • Are we paying for AI baristas nobody talks to? → License waste
  • What custom recipes (agents) are customers ordering? → Agent analytics

The AI-in-One Dashboard is your headquarters monitor — one screen showing exactly who's using AI, how often, which tools, and where the gaps are.

In technical terms: It's a Power BI report that connects to Microsoft Purview audit logs, M365 Admin data, and Microsoft Entra to give you a complete picture of Copilot & Agent adoption across your organisation.

⚠️ Important disclaimer: This is a community/experimental template from Microsoft. Audit logs are the primary data source but are not the sole source of truth for licensing or full-fidelity Copilot usage reporting. For the most accurate insights, Microsoft recommends the M365 Admin Center and Viva Insights. Microsoft support does not cover this template.

Why Should a Customer Care?

Business Question How the Dashboard Answers It
"Are we getting ROI from Copilot licenses?" Shows active vs. inactive licensed users, usage trends over time
"Who should get Copilot next?" Identifies unlicensed users already using free Copilot Chat heavily
"Which departments are lagging behind?" Breaks down adoption by department, location, and role
"Are our custom agents being used?" Agent leaderboard and usage trends show which agents are popular
"Is usage growing or plateauing?" Habit formation charts show weekly active user trends over months
"Where should we focus change management?" Surfaces low-engagement departments and inactive licensed users

Prerequisites & Licensing

What You Need (Before You Start)

# Requirement Details
1 Power BI Desktop Free download from Microsoft Store. You cannot use Power BI Online/Service to open the template — the .pbit file format requires Desktop.
2 Microsoft 365 tenant Any M365 tenant with Copilot licenses deployed
3 Admin access You need specific roles (see table below) to export each data source
4 Copilot usage data Users need to have actually used Copilot — the dashboard reads historical audit logs
5 PowerShell 5.1+ Only needed if you use the automated script method (optional)

Why Power BI Desktop? Why Not Power BI Online?

Aspect Power BI Desktop Power BI Online (Service)
Open .pbit templates ✅ Yes ❌ No — templates can only be opened in Desktop
Load local CSV files ✅ Yes — reads files from your disk ❌ No — can't access local file paths
Edit Power Query ✅ Full editor ⚠️ Limited editing
Free to use ✅ Completely free ⚠️ Requires Pro or PPU license to publish
Bottom line Required for setup Used later for sharing/scheduled refresh

💡 Tip: You set up and load data in Power BI Desktop (free), save as .pbix, then publish to Power BI Service if you want to share with others or set up scheduled refresh.

Admin Roles Required

You need different roles to export each of the 4 data sources:

Data Source Portal Role Required Why
Copilot Audit Logs Microsoft Purview Audit Reader or Compliance Admin Read Copilot interaction audit records
Licensed Users M365 Admin Center Global Admin or Reports Reader Access Copilot usage reports
Agent Inventory M365 Admin Center AI Admin or Global Reader Export the agent registry
Org Data Microsoft Entra User Admin or Global Reader Export user departments, titles, locations

💡 Licensing cost: There is no additional licensing required for the dashboard itself. Power BI Desktop is free, and the data sources are included with your existing M365/Copilot licenses. You only need a Power BI Pro or Premium Per User license if you want to publish and share the report in Power BI Service.

Step-by-Step Setup Guide

Overview of the Process

┌─────────────────────────────────────────────────────────────────┐
│                    SETUP FLOW (5 Steps)                         │
│                                                                 │
│  Step 1          Step 2          Step 3          Step 4         │
│  ┌──────┐       ┌──────┐       ┌──────┐       ┌──────┐        │
│  │Export │       │Export │       │Export │       │Export │        │
│  │Audit  │       │Users  │       │Agents │       │Org    │        │
│  │Logs   │       │List   │       │List   │       │Data   │        │
│  └──┬───┘       └──┬───┘       └──┬───┘       └──┬───┘        │
│     │              │              │              │             │
│     ▼              ▼              ▼              ▼             │
│  📄 CSV          📄 CSV        📄 CSV         📄 CSV          │
│                                                                 │
│     └──────────────┴──────┬───────┴──────────────┘             │
│                           ▼                                     │
│                     Step 5: Load                                │
│                     ┌──────────┐                                │
│                     │ Power BI │                                │
│                     │ Desktop  │                                │
│                     │ Template │                                │
│                     └──────────┘                                │
│                         📊                                      │
└─────────────────────────────────────────────────────────────────┘

Step 0: Install Power BI Desktop

If you don't have Power BI Desktop installed:

  1. Open the Microsoft Store on your Windows PC
  2. Search for "Power BI Desktop"
  3. Click Install (it's free)
  4. Or run this command in PowerShell: 
    winget install --id 9NTXR16HNW1T --source msstore

Step 1: Export Copilot Audit Logs (from Microsoft Purview)

What this file contains: Every time a user interacts with Copilot — whether they ask a question in Word, summarise an email in Outlook, use BizChat, or interact with a custom agent — Microsoft records an audit log entry. This CSV contains every single Copilot interaction in your tenant, including:

  • Who used Copilot (user email)
  • When they used it (timestamp)
  • Which app they used it in (Word, Excel, Teams, Outlook, BizChat, etc.)
  • What type of interaction (Copilot, Agent, Microsoft Agent)
  • Which agent they talked to (if applicable)
  • What AI model was used (GPT-4o, GPT-4o Mini, etc.)
  • What plugins were invoked (Bing search, file search, code interpreter)

How to export it:

  1. Go to security.microsoft.com
  2. Navigate to Audit (in the left sidebar under Solutions)
  3. Click New Search (or use the search form)
  4. Set your date range (e.g., last 90 days for a good overview)
  5. Under Activities - Friendly Names, select:
    • ✅ Copilot Activities – Interacted with Copilot (required)
    • ✅ Copilot Activities – Interacted with a Connected AI App (recommended)
    • ⚠️ Copilot Activities – Interacted with an AI App (optional — requires PAYG billing)
  6. Click Search
  7. Wait for the search to complete (can take 5-60 minutes depending on data volume)
  8. Once completed, click on the search result
  9. Click Export → Download all results
  10. Save the CSV file (e.g., Copilot_Audit_Logs.csv)

💡 Tip: The more days you include, the richer the dashboard will be. 90 days is a good starting point for showing adoption trends. The maximum is 180 days (or 365 days with E5 licensing).

⚠️ Note on "Interacted with an AI App": This activity type captures third-party and custom AI app interactions. It requires pay-as-you-go (PAYG) billing to be enabled on the tenant. If you don't have PAYG, this search will return zero results — that's normal. The dashboard will still work without it.

Step 2: Export Licensed Users List (from M365 Admin Center)

What this file contains: A list of all users in your tenant with a flag showing whether they have a Microsoft 365 Copilot license assigned. The dashboard uses this to:

  • Calculate the license utilisation rate (how many licensed users are actually active)
  • Identify licensed but inactive users (wasting money)
  • Compare licensed vs. unlicensed usage patterns

How to export it:

  1. Go to admin.microsoft.com
  2. Navigate to Reports → Usage
  3. Click on Microsoft 365 Copilot
  4. Open the Readiness tab
  5. In the Copilot Readiness Details table, make sure the column "Has Copilot license assigned" is visible
  6. Click the ⋯ (ellipsis/three dots) menu above the table
  7. Choose Export
  8. Save the CSV file (e.g., M365CopilotUsers.csv)

⚠️ Gotcha — Anonymised user names: By default, M365 Admin Center hides real user names in reports (shows hashed values instead). To fix this:

  1. Go to Admin Center → Settings → Org Settings → Services → Reports
  2. Uncheck "Display concealed user, group, and site names in all reports"
  3. Save — this can take up to 48 hours to take effect
  4. Re-export after the change propagates

Step 3: Export Agent Inventory (from M365 Admin Center)

What this file contains: A registry of all Copilot agents deployed in your tenant — both custom-built agents (from Copilot Studio) and Microsoft's built-in agents. This powers the Agent section of the dashboard, showing:

  • How many agents exist and their types (Declarative, Custom Engine, Copilot Studio)
  • Which agents are published vs. draft
  • Who created each agent and when
  • What capabilities each agent has (OneDrive access, SharePoint access, Graph connectors, etc.)
  • Agent descriptions (use cases)

How to export it:

  1. Go to admin.microsoft.com
  2. Navigate to Settings → Agents (or Copilot → Agents)
  3. Click All Agents
  4. Click Export to Excel (top-right area)
  5. Save the file (it downloads as a CSV despite the button name)
  6. Rename if needed (e.g., Agent365_Inventory.csv)

⚠️ Gotcha — Missing columns: Some versions of the export may be missing columns that the template expects (e.g., "Last updated", "Can read OneDrive and Sharepoint items"). If the dashboard shows a "column not found" error, you may need to add the missing columns manually. See Troubleshooting section below.

Step 4: Export Org Data (from Microsoft Entra)

What this file contains: Organisational information about your users — departments, job titles, office locations, countries, and managers. Without this file, the dashboard can still show Copilot usage, but it cannot break it down by department or role. This is the file that enables:

  • Department-level adoption charts ("Engineering has 85% adoption, Legal has 40%")
  • Role-based insights ("Managers use Copilot 3x more than individual contributors")
  • Location breakdowns ("Auckland office leads in BizChat usage")
  • Manager hierarchy analysis

How to export it:

  1. Go to entra.microsoft.com
  2. Navigate to Identity → Users → All users
  3. Click Download users (top toolbar)
  4. Select the attributes you want (at minimum: UserPrincipalName, DisplayName, Department, JobTitle, City, Country)
  5. Click Download
  6. Save the CSV file (e.g., Org_Data_Entra.csv)

💡 Tip: The richer your Entra data, the better the dashboard insights. If departments and titles are empty in Entra, consider exporting from Viva Insights or an HR system instead. The template needs at minimum a UserPrincipalName and Department column.

Step 5: Load Data into Power BI Template

Now you have 4 CSV files. Let's load them into the dashboard.

  1. Create a folder for your data files (e.g., C:\AIData\) and copy all 4 CSVs there
  2. Open the template file: Double-click AI-in-One Dashboard - 18 03 - csv data sources.pbit
  3. Power BI Desktop will open and show a Parameters dialog with 4 file path fields:
    • Copilot Interactions Audit Logs file path
    • Copilot Licensed Users file path
    • Agents 365 file path
    • Org Data file path
  4. Type the full path to each CSV file (e.g., C:\AIData\Copilot_Audit_Logs.csv)
  5. Click Load
  6. Wait for Power BI to process the data (can take 5-15 minutes for large datasets — the audit log has nested JSON that needs parsing)
  7. Once loaded, save as a .pbix file (File → Save As)

⚠️ Important — Type paths manually! Copy-pasting file paths can introduce invisible characters (BOM, curly quotes) that cause "Illegal characters in path" errors. If you get this error, delete the path and type it manually character by character.

💡 Tip: If Power BI appears frozen during loading, check Task Manager — PBIDesktop.exe will show high CPU if it's still processing. The 70K+ audit log records with nested JSON take time to parse.

Dashboard Tabs — Complete Guide

The AI-in-One Dashboard has 15 tabs organised into 4 sections. Here's what each one shows and how to use it in customer conversations.

Section 1: 🔍 Copilot Overall (Combined View)

These tabs show the big picture across all Copilot surfaces — M365 apps + Copilot Chat + Agents combined.

What it shows How to use it
Total Copilot interactions over time (daily/weekly/monthly) Open with this tab in demos — it's the "executive summary". Shows whether adoption is growing, flat, or declining.
Breakdown by surface (Word, Excel, Teams, Outlook, BizChat, Agents) Identify which Copilot surfaces are most popular. "Your users love Copilot in Teams but barely touch it in Excel — let's talk about Excel training."
Active users trend Track how many unique users are engaging with Copilot each week

🎯 Customer talking point: "This is your Copilot health check at a glance. If the trend line is going up — great, adoption is growing. If it's flat, it's time to invest in change management."

Tab 2: 🔍 Copilot Overall — Combined Leaderboard

What it shows How to use it
Top users by interaction count Identify Copilot champions — power users who can help drive adoption in their teams
Department rankings See which departments are leading and which are lagging
Individual user activity breakdown Drill into a specific user's Copilot usage patterns

🎯 Customer talking point: "Your top 10 users are generating 40% of all Copilot interactions. These are your internal champions — empower them to train their peers."

Section 2: 📈 M365 Copilot (Licensed Users)

These tabs focus specifically on licensed Copilot users using Copilot within M365 apps (Word, Excel, PowerPoint, Outlook, Teams, BizChat).

What it shows How to use it
Licensed user activity trends over time Track adoption trajectory for paid Copilot users
App-by-app breakdown (Word, Excel, PPT, Outlook, Teams, BizChat) Identify which M365 apps see the most Copilot usage
Daily/weekly active user counts Measure engagement intensity

🎯 Customer talking point: "You've deployed 200 Copilot licenses. Are they being used? This tab shows that 150 are active weekly — that's 75% adoption. Industry average is around 50%, so you're ahead of the curve."

Tab 4: 📈 M365 Copilot — Leaderboard

What it shows How to use it
Rankings of licensed users by Copilot usage Find your power users and low adopters
Filterable by department, location, role Slice and dice to find patterns

Tab 5: 📈 M365 Copilot — Habit Formation

What it shows How to use it
Weekly active user progression over weeks/months The key metric for long-term success — are users forming a Copilot habit?
Retention curves Shows if users who start using Copilot continue using it
Consistency patterns Identifies users who use Copilot every day vs. sporadically

🎯 Customer talking point: "Adoption is one thing, but habit formation is what drives ROI. This chart shows that after 8 weeks, 60% of your users are still using Copilot weekly — that's a strong signal that it's adding value to their workflow."

Section 3: 📊 Copilot Chat (Web) — Unlicensed Users

These tabs focus on users accessing Copilot Chat via the web (copilot.microsoft.com) — including unlicensed users who use the free tier.

What it shows How to use it
Free Copilot Chat usage over time Track organic demand from unlicensed users
Volume of interactions from non-licensed users Quantify the "shadow AI" usage in your org

🎯 Customer talking point: "Look at this — 25 unlicensed users are already using Copilot Chat 3-5 times per day. They've found value on their own. These are your best candidates for paid Copilot licenses because you know they'll use it."

Tab 7: 📊 Chat (Web) — Leaderboard

What it shows How to use it
Top unlicensed Copilot Chat users ranked by activity License recommendation goldmine — sort by usage to find who deserves a license
Department and role information Prioritise license allocation by business impact

🎯 Customer talking point: "Instead of guessing who should get the next batch of Copilot licenses, use data. This leaderboard shows the 20 unlicensed users with the most Copilot Chat activity. Deploy licenses to them first — guaranteed adoption."

Tab 8: 📊 Chat (Web) — Habit Formation

What it shows How to use it
Consistency of free Copilot Chat usage Shows which unlicensed users are regular users vs. one-time trialists

Tab 9: 📊 License Prioritisation

What it shows How to use it
Combined view of licensed vs. unlicensed usage The decision-making tab — see who has a license and isn't using it, and who doesn't have a license but should
License waste identification Find licensed users with zero or very low activity
License opportunity identification Find unlicensed users with high free-tier activity

🎯 Customer talking point: "This tab tells you two things: (1) You have 15 licensed users who haven't touched Copilot in 60 days — consider reassigning those licenses. (2) You have 8 unlicensed users averaging 5+ daily interactions on the free tier — they're proving the value case on their own."

Section 4: 🤖 Agents

These tabs focus on Copilot agents — custom-built AI assistants created in Copilot Studio or declaratively.

What it shows How to use it
Agent interaction volume over time Track if custom agent adoption is growing
Which agents are being used See which custom agents are popular vs. ignored

🎯 Customer talking point: "You've built 12 custom agents but only 4 are getting regular usage. Let's look at what makes those 4 successful and apply those patterns to the others."

Tab 11: 🤖 Agents — Leaderboard

What it shows How to use it
Top agents ranked by interaction count Identify your most valuable agents
Top users of agents Find your agent power users
Agent type breakdown (Declarative, Custom Engine, Copilot Studio) Understand the agent landscape

Tab 12: 🤖 Agents — Habit Formation

What it shows How to use it
Weekly agent usage consistency Are users returning to agents regularly or was it a one-time trial?

Tab 13: 🤖 Agents — Health Check

What it shows How to use it
Agent inventory overview How many agents are published vs. draft
Agent capabilities matrix Which agents have access to OneDrive, SharePoint, Graph connectors
Agent freshness When agents were last updated

🎯 Customer talking point: "You have 32 agents registered but 8 are still in draft and 5 haven't been updated in 3+ months. Let's do an agent health check — retire the unused ones and promote the successful ones."

Tab 14: 🤖 Agents — Use Cases

What it shows How to use it
Agent descriptions and use-case summaries Understand what each agent was built for
Agent categorisation by department/function See the spread of agents across business areas
Agent capability details Which agents can read files, generate images, use code interpreter

🎯 Customer talking point: "This tab shows your agent portfolio. You've got IT help desk, HR policy, sales proposals, contract review — these are real business process automations. The next step is identifying 3-5 more high-impact use cases."

Section 5: 📖 Reference

Tab 15: 📖 Metric Glossary & Guide

What it shows How to use it
Definitions of all metrics used in the dashboard Reference guide for understanding what each number means
Data source explanations Where each piece of data comes from
Methodology notes How calculations are performed

Using the Dashboard for Customer Conversations

Start Here                    Deep Dives (pick 1-2)           Close
┌──────────┐    ┌──────────────────────────────────────┐    ┌──────────┐
│ Combined │    │  📈 M365 Copilot Trends & Habits    │    │ License  │
│ Trends   │───▶│  📊 Chat Web Leaderboard            │───▶│ Priority │
│ (Tab 1)  │    │  🤖 Agent Usage & Health            │    │ (Tab 9)  │
└──────────┘    └──────────────────────────────────────┘    └──────────┘
  "Here's        "Let's drill into                         "Here's your
  the big         the details..."                           action plan"
  picture"
  1. Open with Tab 1 (Combined Trends) — "Here's your overall Copilot health"
  2. Drill into 1-2 areas based on what the customer cares about:
    • If they care about ROI → M365 Copilot Habit Formation (Tab 5)
    • If they care about license spend → License Prioritisation (Tab 9)
    • If they care about agents → Agent Leaderboard + Use Cases (Tabs 11 & 14)
    • If they care about expanding licenses → Chat Web Leaderboard (Tab 7)
  3. Close with Tab 9 (License Prioritisation) — "Here's your data-driven action plan"

Key Stories to Tell Customers

Story Tabs to Use Message
"You're getting ROI" Tabs 1, 3, 5 Show growing trend + habit formation = sustained value
"Expand smartly" Tabs 6, 7, 9 Show unlicensed users with high free-tier usage = guaranteed adoption
"Reduce waste" Tabs 4, 9 Show licensed users with zero activity = reassign those licenses
"Agents are the future" Tabs 10, 11, 14 Show agent usage growing + agent use cases = business process automation
"Champion your champions" Tabs 2, 4 Show top users = train-the-trainer candidates

Architecture Overview

┌──────────────────────────────────────────────────────────────────┐
│                     DATA SOURCES                                 │
│                                                                  │
│  ┌─────────────────┐  ┌──────────────────┐  ┌────────────────┐  │
│  │ Microsoft Purview│  │  M365 Admin      │  │  Microsoft     │  │
│  │ (Audit Logs)    │  │  Center          │  │  Entra ID      │  │
│  │                 │  │                  │  │                │  │
│  │ Who used Copilot│  │ Who has a license│  │ Department,    │  │
│  │ When, which app │  │ License status   │  │ title, city    │  │
│  │ Which agent     │  │ Agent registry   │  │ manager        │  │
│  └────────┬────────┘  └───────┬──────────┘  └───────┬────────┘  │
│           │                   │                      │           │
└───────────┼───────────────────┼──────────────────────┼───────────┘
            │                   │                      │
            ▼                   ▼                      ▼
┌──────────────────────────────────────────────────────────────────┐
│                      4 CSV FILES                                 │
│                                                                  │
│  📄 Copilot_Audit_Logs.csv  📄 M365CopilotUsers.csv            │
│  📄 Agent365_Inventory.csv  📄 Org_Data_Entra.csv               │
└──────────────────────┬───────────────────────────────────────────┘
                       │
                       ▼
┌──────────────────────────────────────────────────────────────────┐
│                  POWER BI DESKTOP                                │
│                                                                  │
│  .pbit template → Power Query parses JSON → 15-tab dashboard    │
│                                                                  │
│  Save as .pbix → Publish to Power BI Service → Share & Refresh  │
└──────────────────────────────────────────────────────────────────┘

Troubleshooting & Gotchas

Issue Solution
UPNs hashed in licensed users export Admin Center → Settings → Org Settings → Services → Reports → uncheck "Conceal user names". Takes up to 48 hours.
"Supplied file path must be a valid absolute path" Don't copy-paste paths. Type them manually to avoid invisible characters.
"Illegal characters in path" Caused by curly quotes or BOM from clipboard. Delete and retype the path.
Agent 365 "column not found" error Template expects columns like "Last updated", "Can read OneDrive and Sharepoint items". Add missing columns via PowerShell or re-export from a newer Admin Center version.
"Interacted with an AI App" returns 0 results This audit activity requires pay-as-you-go (PAYG) billing enabled on the tenant. The dashboard works without it.
Dashboard takes 10+ minutes to load Normal for large datasets. The Power Query parser needs to parse nested JSON in every audit log record.
Power BI Desktop not in winget Install from Microsoft Store: winget install --id 9NTXR16HNW1T --source msstore
Power BI appears frozen during load Check Task Manager — high CPU on PBIDesktop.exe means it's still processing. Be patient.

Additional Resources

Resource Description
GitHub Repository Source code, scripts, and template files
AI-in-One - Interpretation Guide.pdf Official guide on how to read and interpret the dashboard metrics
AIinOne_Storyboard_2601.pptx Dashboard design storyboard from the Microsoft team
scripts/ folder PowerShell scripts for automated data export via Graph API
scripts/automation/ Azure Automation runbooks for scheduled, unattended data refresh

What's Next After Setup?

  1. Share with stakeholders — Publish to Power BI Service and share with your Copilot adoption team
  2. Set up scheduled refresh — Use the SharePoint data source template or Azure Automation to auto-refresh weekly
  3. Customise for your org — Add your company logo, adjust date ranges, add custom visuals
  4. Track over time — Refresh monthly to show adoption trends in quarterly business reviews
  5. Combine with Viva Insights — For deeper sentiment and productivity analytics beyond usage counts

Progress Log

Date What happened
2026-03-29 Idea added to vault, project plan created
2026-03-29 Phase 1 ✅ — Lab tenant verified: 24/25 Copilot licenses, E5, Global Admin
2026-03-29 Phase 2 ✅ — All 4 data sources exported via PowerShell + Graph API
2026-03-29 Phase 3 ✅ — Power BI template loaded with data! Dashboard is live
2026-03-29 Mock data generated: 71,952 interactions, 220 users, 6 months
2026-03-29 Comprehensive setup & user guide documentation completed