Everything you need to know about Copilot reporting.

Microsoft 365 Copilot adoption is best measured across five layers: license assignment (who has it), enablement (who has signed in), usage (who is actively prompting), depth (how many surfaces — Word, Excel, Teams, Outlook — they use), and sentiment (do they say it's helping).

The AI-in-One Dashboard combines all five into a single Power BI report by pulling from Microsoft 365 admin center exports, Viva Insights Copilot Dashboard data, and optional Purview audit logs.

No — reporting is the diagnostic, not the cure. Dashboards tell you where you are; change management and culture are what move you forward. Without a deliberate programme behind the numbers, the metrics plateau no matter how good the dashboards are.

Across the customers we work with, the biggest adoption lifts come from:

• Leadership visibly using Copilot — modelling the behaviour you want to see
• Role-specific enablement workshops — finance, HR, sales, engineering each have different prompts
• Internal champions who answer "how do I…" questions in real time
• Shared prompt libraries so good prompts travel across teams
• Recognition for early experimenters — make it socially rewarding to try things

The Analytics Hub gives you the measurements. Pair them with a deliberate change-management and culture programme and both curves bend upward together.

Industry-wide Copilot benchmarks sound useful — and they're the most-asked-for thing this team gets — but the honest answer is that cross-company benchmarks are usually misleading rather than helpful. Copilot adoption is shaped by so many context-specific variables that a peer's number rarely maps to your reality.

The factors that actually move the adoption curve include:

• Organisational culture — collaborative vs siloed, experimental vs conservative
• Company size and structure — a 5,000-person flat org and a 50,000-person matrix behave nothing alike
• Industry — regulated industries adopt slower; knowledge-work industries accelerate fast
• Features enabled or disabled — BizChat, agents, image generation, plugins each shift behaviour
• Rollout strategy — big-bang, cohort waves, or role-based
• Change-management investment — one all-hands vs a sustained enablement programme
• Internal AI promotional activities — champion networks, prompt libraries, internal learning days, recognition programmes

A competitor's "30% weekly active users" might be a 500-seat regulated-bank pilot, or a 50,000-seat tech company three years into the journey. Same number, completely different meaning.

The benchmark that actually defends decisions is yourself. Track month-over-month change against your own baseline, and progress against an internal hypothesis — e.g. "if our champion programme works, we expect weekly-active users to climb five points by Q3". That's the comparison that holds up in front of a CFO.

There isn't one best template — the right report depends on the question. Use this matchup:

Executive monthly review → AI-in-One Dashboard
Prompt-level chat analytics → Chat Intelligence Report
Declarative and custom agent telemetry → Agent Intelligence Report
ROI and license utilization defence → Super User Impact + Decoding Super Usage
Sentiment and adoption psychology → Copilot Adoption & Sentiment Report

The Explore Reports page maps every business question to the right template in one click.

The Viva Insights Copilot Dashboard is the native Microsoft dashboard included with Viva Insights — it shows adoption, behavioural change, and ROI for tenants with the right Viva licensing.

The Analytics Hub is an open-source toolkit from the same Microsoft Copilot ROI Advisory team that extends Viva data with prompt-level intelligence, agent telemetry, super-user analysis, sentiment, and PowerShell exporters — for tenants that need to combine multiple data sources, or that don't have Viva at all.

See the full side-by-side comparison.

License utilization is measured by comparing assigned licenses against active, recently active, and depth-of-use cohorts.

The Super User Impact and Decoding Super Usage reports surface utilization by cohort, identify dormant licenses for reclamation, and show which licensed users are driving the majority of measurable value — exactly the data leadership needs to defend or reallocate license waves.

Yes. The Agent Intelligence Report ingests Copilot agent telemetry — declarative agents, custom engine agents, and Studio-built agents — and breaks down usage by agent, by author, by department, and by outcome.

It pairs with the Chat Intelligence Report so leadership can see both human chat patterns and agent invocations in a single Power BI model.

No. Most reports in the Analytics Hub run on data you already have — Microsoft 365 admin center CSV exports, Purview audit logs, or the included PowerShell exporters such as PAX.

A subset of reports lights up additional pages when Viva Insights Copilot Dashboard data is available, but the core adoption, usage, agent, and sentiment reporting works without a Viva license. The Data Sources map shows which template needs what.

Purview Audit retains Copilot interactions for up to a year (with the right SKU). PAX — the PowerShell Audit Exporter included with the Analytics Hub — automates the search, retrieval, and CSV export of those audit records into a shape the Chat Intelligence and Agent Intelligence reports can consume.

Setup is documented on the Explore Reports page. The only prerequisites are Purview Audit and an account with Audit Reader + Compliance Administrator rights.

Yes. Every report, template, and exporter in the Analytics Hub is published under the MIT license on the microsoft/Analytics-Hub GitHub repository.

There is no licensing cost, no telemetry calls back to Microsoft from the templates, and no Microsoft consulting engagement required to run them.

The Microsoft Copilot ROI Advisory Team — a group of People Scientists, Business Value Advisors, Customer Success Managers, and Solutions Architects inside Microsoft who work daily with the largest Copilot rollouts.

Every report originated from a real customer engagement and has been pressure-tested in production tenants before it ships here.

Three options, in order of speed:

1. Use the Feedback page to email the team directly (problem, suggestion, or praise)
2. Open an issue on the microsoft/Analytics-Hub GitHub repository
3. Check the troubleshooting section in the report's own README, plus the admin-friendly email templates you can forward to your IT or licensing team