AI metrics
The following metrics are used in the AI activity dashboards.
Adoption score
The Adoption Score quantifies the overall success of the AI tool integration across your organization. It is a weighted metric that prioritizes two primary indicators—Seat Utilization and User Engagement—while also factoring in Code Acceptance Rates and Feature Adoption.
A higher score reflects a more successful rollout. For easy interpretation, the score is color-coded by position on a 0–100 scale, with Excellent, Good, Fair, and Poor categories.
It is composed of the following four metrics:
Seat utilization
Seat utilization = (active users / total seats) x 100.
An active user is a user who has used an AI tool during the current billing cycle.
This metric identifies active users who have engaged with an AI tool—whether by receiving or accepting suggestions—and compares them against the total number of currently assigned seats.
By analyzing these figures alongside seat utilization, leadership can easily identify underused licenses to better allocate spending or reassign excess seats to other team members.
Rating ranges:
Poor: <40
Fair: 40-59
Good: 60-79
Excellent: 80-100
Data source:
For Copilot: Data is pulled directly from the GitHub Copilot Org Metrics endpoint for active users over the date range specified at the top of the page (currently 90 days).
Code acceptance rate
Code acceptance rate = (total acceptances / total suggestions) x 100.
This metric tracks the ratio of accepted suggestions to total completions offered, providing a clear view of how much AI-generated content is actually used.
By monitoring this rate, you can determine if suggestions are relevant and if your teams trust the tool’s output, helping to confirm whether a drop in the metric correlates with a perceived decline in AI quality.
Rating ranges:
Poor: 0-9 or 76-100
Fair: 10-15 or 55-75
Good: 16-25 or 36-54
Excellent: 26-35
The optimal range is between 26% and 35%, which indicates healthy AI-assisted development.
Data source:
For Copilot: Copilot provides a number of suggestions and a number of acceptances per day through its code completion metrics.
User engagement
User engagement rate = (average daily engaged users / average daily active users) x 100.
This metric tracks the average daily interactions per user, including requests submitted, suggestions accepted, and chat features used.
By monitoring these activity patterns, you can determine whether your adoption is truly active or merely passive, helping you distinguish between teams that merely hold licenses and those that are deeply engaged with AI features.
Rating ranges:
Poor: <40
Fair: 40-59
Good: 60-79
Excellent: 80-100
Data source:
For Copilot: Copilot's daily metrics distinguish between active and engaged users. Data is pulled from the Copilot’s API.
Feature adoption
Chat usage rate = (average daily active chat users / average daily active users) x 100.
This metric tracks the usage rate of advanced AI tool features, such as chat-based code explanations and debugging assistance, among your active users. It measures the average daily number of developers engaging with these tools to solve complex development tasks.
While moderate usage indicates healthy adoption of advanced capabilities, extremely high rates are monitored to identify potential overdependence on chat for basic reasoning.
Rating ranges:
Poor: 0-9
Fair: 10-14 or 75-100
Good: 15-24
Excellent: 25-74
Moderate usage indicates healthy adoption of advanced AI features.
Total seats
Utilization rate = (active users/total seats) x 100
This metric displays the total number of assigned seats based on current allocations rather than a historical range. When used alongside seat utilization data, it allows leadership to identify underused licenses, optimize spending, and reassign excess seats to other team members.
Data source:
For Copilot: Data is pulled directly from the GitHub Copilot Org Metrics endpoint for active users over the date range specified at the top of the page (currently 90 days).
For Aggregate: Total number of aggregated seats provisioned in your organization.
Performance score
This balanced metric measures overall engagement by weighting chat interactions and code acceptance quality.
By equally balancing chat engagement (50%) and code acceptance (50%), the score provides a comprehensive view of AI adoption without relying on arbitrary volume thresholds.
Daily average suggestions
Daily average = total suggestions + number of days
This metric tracks the mean volume of code-completion suggestions provided by the AI tool or aggregated tools each day over the chosen timeframe.
A higher daily average indicates consistent developer engagement with AI features across the team.
Weekly growth
Growth rate = Average of all [(weekn - weekn-1) / weekn-1)] x 100.
This metric identifies usage trends by calculating the average week-over-week growth across all available data points. The reporting period is divided into 7-day increments, with any remaining days treated as a partial week.
Example with 25 days:
Week 1: Days 1–7
Week 2: Days 8–14
Week 3: Days 15–21
Week 4: Days 22–25 (Partial)
Growth rates are determined between each consecutive week and then averaged. A positive result indicates increasing AI adoption across your team.
Peak day
This metric identifies the single 24-hour period with the highest volume of AI tool or aggregated suggestions. It serves as a benchmark for when your developers are most heavily engaged with AI assistance.
Peak days typically coincide with major project deadlines, new feature development, or collaborative team coding sessions. Analyzing these surges helps you understand the specific contexts where Cursor provides the most value to your workflow.