Licenses are active. The internal announcement has gone out. And three months later, fewer than 4 in 10 employees with access to Copilot are actually using it. This is not an exception: it is the norm. According to data from Recon Analytics gathered by multiple analysts, only 35.8% of employees with active access to Copilot actually use it. Copilot adoption does not fail because of the product; it fails due to the lack of a measurement system and an acceleration plan that works in production.
This article gives you exactly that: the KPIs that matter, how to read them from the admin center, concrete prompts by role, an actionable governance checklist, and the levers to move from a 20% activation rate to 70% or more in 90 days.
[[NSEO_CTA:2]}
Discover how LATAM companies accelerate adoption with AI.Sign up and see real cases with metrics.[[/NSEO_CTA]]
The real problem: assigned licenses ≠ real adoption
There is a gap that very few executives want to see on their dashboard. In many organizations, fewer than 40% of employees with access to Copilot use it actively. The remaining majority generates cost without producing results. Worse still: even organizations that invest heavily in licenses see adoption stall weeks after deployment. The tool gets tried, does not seem immediately transformative, and people return to their usual workflows.
The cause is not technical. The problem is not that Copilot is not integrated into the tools the team already uses every day. The problem is one of training and change management. In other words: access does not automatically equal adoption. And without a metrics system to make it visible, the problem goes unnoticed until someone asks about ROI.
A fact worth keeping in mind when talking about licenses: E3 licenses include free Copilot Chat, but full Copilot—the one that reads your emails, files and meetings via Microsoft Graph—requires the paid add-on, available on top of E3, E5 or Business Premium. Many organizations confuse both levels and overestimate their actual user base.
The Copilot adoption KPIs that really matter
Before talking about acceleration, you need to know what to measure. Most IT teams only monitor assigned licenses and initiated sessions. That is noise, not signal. The KPIs that enable decision-making are different.
Activation rate and weekly active users
The activation rate measures the percentage of assigned licenses with at least one real action in the last 28 days. It is the first thermometer. The Copilot usage report in the Microsoft 365 admin center answers the key question: “We assigned Copilot licenses—are people actually using them?”
But one-time activation is not enough. What distinguishes a successful adoption project from a cosmetic one is sustained frequency. Microsoft classifies enterprise users into four groups: power users, regular users, beginners and non-users. Companies no longer just look at the number of adopters, but at the consistency and frequency of use. Power users are those who perform Copilot actions at a high weekly frequency consistently over several weeks.
Microsoft Adoption Score and the 3-day threshold
The Microsoft Adoption Score is a 100-point indicator available in the Admin Center that measures how thoroughly your organization has integrated M365 Copilot into daily work. The score is calculated by measuring how often licensed users interact with Copilot at least three days per week on average, because users who exceed that threshold are significantly more likely to become consistent long-term adopters. If all your licensed users reach that threshold, your AI Adoption Score is 100.
Depth of use by application
Not all activity is equal. A user who only uses Copilot to summarize meetings in Teams contributes less value than one who integrates it into Outlook, Word and Excel. M365 Copilot administrators can track active users and prompt counts through the admin center, gaining visibility into real usage versus provisioned seats. Organizations that measure internally focus ROI on three metrics: hours saved per user per week, improvement in task completion rate, and reduction in context switching between applications.
This breakdown by application is critical for LATAM, where workflows in manufacturing, retail and financial services have very different usage patterns. A depth KPI by role and by application tells you where there is real value and where there is hidden friction.
90-day retention ratio
The most honest indicator of sustained adoption is how many users who started using Copilot in month 1 are still using it in month 3. Microsoft recommends a three-phase ROI measurement approach: baseline productivity metrics before deployment, 90-day activation tracking, and 6-month outcome surveys measuring time saved per task in the most frequent use cases by role. If retention drops before 90 days, there is a training or use-case relevance problem, not a technology problem.
How to access the admin center and read the adoption report
Knowing that the KPIs exist is not enough: you need the exact path to reach them. Here is the step-by-step access for the technical profile managing the deployment.
Path in the Admin Center (basic reports)
According to Microsoft’s official documentation, access to Copilot reports follows this path: sign in to the Microsoft 365 Admin Center as AI Administrator or Global Administrator → go to Reports → Usage → select Microsoft 365 Copilot → Readiness or Usage. The key columns you should read are:
- Copilot licenses assigned: licenses assigned in the tenant (includes disabled mailboxes; keep this in mind to avoid inflating the denominator).
- Active Copilot users (28 days): users who performed at least one intentional action with Copilot in Teams, Outlook, Word, Excel, PowerPoint, OneNote or Copilot Chat in the last 28 days. This is your real activation rate.
- App integration: breakdown by application; tells you where usage is concentrated and where there is friction.
- Top agents: most-used Copilot agents; useful for detecting whether teams are already extending usage toward automation.
Path in Viva Insights (advanced analysis)
For deeper analysis, the Copilot Dashboard in Viva Insights is available to any customer with an M365 or O365 enterprise subscription and an active Exchange Online account, without needing an additional paid Viva Insights license. The path is: open the Viva Insights app in Teams → select Copilot Dashboard. From there you access the Readiness, Adoption, Impact and Sentiment pages. Data covers the last 28 days with a maximum delay of 6 days.
For advanced analysis with Power BI, go to Viva Insights → Analysis → Templates → Microsoft 365 Copilot adoption. The Copilot adoption report in Viva Insights includes a Power Users tab that allows you to identify the teams with the highest usage frequency and how they integrate it into their daily work. Historical Copilot data is available for up to 13 months.
Sample prompts by role: the most actionable content for your champions
Prompt engineering—the ability to formulate effective instructions for Copilot—is the most underestimated gap in adoption projects. Most employees have never received training on how to interact with an AI model. Vague prompts produce vague results, frustration arrives quickly and the user gives up. Here are role-validated prompts to use from day one with your champions.
| Role | Application | Sample prompt |
|---|---|---|
| Finance | Excel | “Analyze this Q1 expense table by department and highlight the three line items with the greatest variance from budget. Generate a 5-line executive summary.” |
| Finance | Outlook | “Draft an email to the accounting team requesting the closure of pending invoices before Friday. Professional tone, maximum 3 paragraphs.” |
| Sales | Teams | “Summarize the commitments and next steps agreed in today’s client meeting. Format as an action list with owner and deadline.” |
| Sales | Word | “Create a draft commercial proposal for [client name] based on the pain points identified in the last call. Include an estimated ROI section.” |
| HR | Word | “Write a job description for a Senior Data Analyst. Include responsibilities, technical requirements and soft skills. Inclusive tone.” |
| HR | Outlook | “Summarize the last 10 emails in this onboarding thread and extract the unresolved questions. Prepare a draft reply that addresses all of them.” |
| Operations | Teams | “Identify the three bottlenecks mentioned in the last four project status meetings and suggest corrective actions for each one.” |
Share this table with your champions in the expansion phase. A concrete, functional prompt has more contagion power than any corporate communication. The Role-Goal-Context-Expectations format (who Copilot should be, what you want, relevant context, expected format) is the framework that consistently produces the best results.
How to diagnose resistance to change before it stalls the project
Resistance to change is not always visible. It is not that users say “I don’t want to use this.” It is that they keep doing things the way they always have because no one has demonstrated to them that Copilot saves them time in their specific work. There are three resistance patterns we see repeatedly in LATAM projects.
Resistance due to distrust in results
44.2% of users who abandoned Copilot cite distrust in the responses as the main reason. This is not resolved with more internal communication; it is resolved with use cases validated in the organization’s real context. A finance user in Mexico or Colombia needs to see that Copilot works with their data, in their processes, before trusting it.
Resistance due to an unmanaged learning curve
Many organizations underestimate the learning curve associated with AI tools. The classic mistake: launching Copilot with a 30-minute webinar and assuming people already know how to use it. Training must be role-based, with concrete business use cases, not generic product tutorials. Vague prompts produce vague results; without training in prompt engineering, frustration arrives before value does.
Structural resistance: no executive sponsor
Organizations that exceed a 50% activation rate in the first 90 days typically deploy three enablement tactics: executive sponsorship, role-specific workflow guides and a dedicated Copilot community channel for sharing prompts and use cases. Without an executive who leads by example and communicates urgency, the project becomes an IT initiative that no one feels ownership of.
Real LATAM case: from 22% to 74% activation in 11 weeks
A financial services company headquartered in Colombia with approximately 800 employees came to us with a classic problem: it had been running active M365 Copilot licenses for four months with a 22% activation rate. The diagnosis revealed three factors: absence of visible executive sponsorship, generic product training (not role-based) and unaudited SharePoint permissions that were generating distrust in the compliance team.
The intervention plan followed the 30/60/90-day model described in this article. In the first four weeks, 12 power users were identified in the risk, treasury and customer service areas, and the five use cases with the highest volume of time invested were validated. In week five, a director of operations was brought in as executive sponsor with weekly communication of results. At eleven weeks, the activation rate had risen to 74% and the average time to produce risk reports had been reduced by 31%. The decisive factor was not the technology; it was the combination of validated use cases, role-specific training and an executive champion who shared their own prompts in the project’s Teams channel.
The 30/60/90-day acceleration plan
The speed of adoption is not random. Organizations that reach 70% active usage in 90 days follow a recognizable pattern. Here is the roadmap we apply with our clients.
Days 1–30: pilot with power users
Successful Copilot deployments follow a three-phase model: pilot with 50–100 power users, expansion to business unit champions, and broad rollout with role-specific training. Each phase is validated with metrics from the Microsoft 365 admin center.
In the pilot phase, the goal is not to scale; it is to identify the 3–5 use cases with the greatest measurable impact in your organization. Meetings in Teams, drafts in Outlook, analysis in Excel: choose the highest-volume workflows and document time saved before and after. That data is your argument for expansion.
Days 31–60: expansion to champions by area
With validated use cases, the next step is to designate champions in each functional area: finance, sales, operations, HR. Their role is not technical support; it is demonstrating value in the language of each team. A finance champion in a LATAM manufacturing company has more credibility to convince their colleagues than any corporate communication.
During this phase, usage KPIs must be monitored weekly. The Copilot adoption report allows you to answer questions such as: How are active users evolving week by week? Is growth cross-cutting or driven by specific groups? Are there sustained increases following enablement actions such as training sessions or champion campaigns? Use the Microsoft Adoption Score as a weekly traffic light: if the score does not rise after an enablement action, review the use case or the training—do not wait.
Days 61–90: broad rollout and ROI measurement
In the final phase, the focus shifts from adoption to business outcomes. The Forrester TEI study for a composite organization of 25,000 employees estimates an ROI of 116%, an NPV of $19.7 million and 9 hours saved per user per month, as well as a 25% reduction in new employee onboarding time. Consult the official Forrester study for figures applicable to your company’s size and profile.
These are the benchmarks you should bring to your steering committee. Not as promises, but as references to compare against your organization’s real data. The difference between an AI project that survives its first year and one that gets cancelled lies in whether you can show these numbers with your own data.
ROI metrics that convince a CEO or CFO
Technical teams talk about activation rates and prompts per week. CEOs and CFOs talk about cost per hour saved, reduction in time on critical processes and return on investment in licenses. You need both languages.
The four ROI metrics that work in executive presentations are:
- Hours saved per user per week: measures the direct impact on productivity. Public studies show between 14 and 26 minutes saved per day per user. Multiply by the average hourly cost of your workforce and you have the return in dollars or euros.
- Reduction in meeting time: organizations using Copilot in Teams report between 15% and 25% greater productivity in meeting-related activities, and employees who save an average of 25 hours per year in preparation and follow-up.
- Active user retention rate at 90 days: this is the indicator that value is sustained, not just perceived in the first month of novelty. The reference benchmark is achieving active user rates of 70–80% among enabled users over a six-month tracking period.
- Cost per inactive license: the most uncomfortable number, but the most persuasive. With the Copilot add-on at $30 per user per month, an organization with 500 licenses and 65% inactivity is paying more than $9,750 per month for nothing. Quantifying this turns adoption into a financial urgency, not just a technological one.
Governance and data: actionable checklist before deployment
There is an obstacle that appears in almost every Copilot adoption project in LATAM and that IT teams usually discover too late: data governance. Copilot does not create new data access paths; it works with the permissions that already exist. The most important security variable is not the tool itself, but whether your organization’s permissions and governance are correctly configured before deployment. In markets such as Mexico, Colombia or Brazil, where regulatory compliance (LGPD, Habeas Data) adds layers of complexity, this step is not optional.
Use this checklist before scaling beyond the pilot:
- SharePoint permissions audit: review sites with “Everyone” or “Everyone except external users” access and restrict those containing sensitive information.
- Microsoft 365 groups review: remove obsolete groups and verify that membership reflects the current organizational structure.
- Sensitivity Labels: configure labels in Microsoft Purview to classify confidential documents; Copilot respects these labels when generating content.
- Review of shared OneDrive files: identify files shared externally that should not be accessible from Copilot.
- Data retention policy: define which data can be queried by Copilot and how long prompt logs are retained.
- LGPD compliance (Brazil): verify that the processing of personal data in prompts complies with the legal bases defined in your RIPD.
- Habeas Data compliance (Colombia/others): ensure that employee and customer data processed by Copilot has the corresponding authorization.
- Review of Copilot connectors and plugins: audit which third-party connectors are enabled in Copilot Studio for your tenant.
- Microsoft Purview Audit configuration: activate Copilot activity logging to be able to audit which prompts are executed and on what data.
- Employee communication on privacy: clearly inform employees what data Copilot uses, whether prompts are used to train models (currently Microsoft does not use M365 Copilot prompts to train base models) and how to exercise access rights.
- Data incident response plan: define the protocol if Copilot exposes sensitive information to an unauthorized user.
- Quarterly permissions review: governance is not a one-time event; schedule a review every 90 days aligned with the adoption cycle.
Real benchmarks: where does your organization stand?
To know whether your adoption speed is good, you need a point of comparison. The Copilot Dashboard in Viva Insights allows you to compare your organization with similar cohorts by manager type, region and job function, expressing the result as a percentage of active Copilot users. But beyond the native tool, these are the ranges we observe in the market:
| Profile | Typical activation rate | Differentiating factor |
|---|---|---|
| Large company with dedicated transformation team | 55–70% | Executive sponsorship + role-based training |
| Mid-sized company (200–999 employees) | ~28% average activation | Limited by scarcity of IT resources and internal champions |
| Deployment without an adoption plan | 15–25% | Licenses only; no training or champions |
| Implementation with workflow integration and change management | 85–95% monthly active usage | Specific use cases + prior governance |
Want to go beyond base Copilot and explore how autonomous agents can automate complete workflows in your organization? Discover how Copilot Studio and Microsoft 365 agents work to extend the value of your current licenses without the need for new integrations.
Copilot adoption is not an IT project: it is an organizational change project with a technology layer. Organizations that understand it this way—measuring with the right KPIs, training by role, managing governance from day one and scaling with champions—are the ones that turn license spending into real competitive advantage. The first step is knowing exactly where you stand today: open the Admin Center, go to Reports → Usage → Microsoft 365 Copilot and read your real activation rate. Everything else starts there.
Sources
- Microsoft Copilot Adoption Rate and Statistics in 2026: Usage Data and Trends
- The Definitive Microsoft 365 Copilot Adoption Guide for Businesses [2026]
- Microsoft Copilot Enterprise Adoption in 2026: What the Data Shows
- Microsoft Copilot Statistics for 2026: Usage and Trends
- Microsoft Copilot Statistics 2026: Users, Growth, and ROI • XtendedView