Enterprise AI Measurement Guide
Weekly Pulse
AI Adoption Stats
What does this week's AI-Assisted PR rate and Heavy/Medium/NA usage split show about AI coding adoption?
What it shows
The AI Adoption stats block shows two adoption signals for the current week: AI-Assisted PRs % (the share of pull requests that week that involved AI coding tool assistance) and the AI Usage split (counts of engineers classified as Heavy, Medium, or NA users based on their activity level). Together these two numbers give engineering leaders a weekly read on both output penetration and workforce segmentation.
Why it matters
AI-Assisted PRs % is a more meaningful weekly adoption signal than WAU because it connects directly to output. A week where 68% of PRs were AI-assisted is a different statement than "68% of engineers logged in", it means AI showed up in more than two-thirds of the code that shipped that week. The Heavy/Medium/NA split surfaces the adoption tier distribution: an org with 80% WAU but 60% NA classification has engineers who logged in but generated no meaningful AI activity, a nominal adoption problem that needs different intervention than a genuine non-adoption problem.
The Larridin angle
The weekly cadence of the AI Adoption stats card makes trend detection fast. A week where NA count rises from 4 to 9 engineers is a signal to investigate, whether it reflects incident load pulling engineers away from development work, a tool outage, or genuine disengagement. Larridin's weekly resolution on this metric is what makes those signals actionable before they accumulate into a monthly trend that requires retrospective explanation.
Related Weekly Pulse Metrics
Common questions
What is the significance of AI-Assisted PRs % in the AI Adoption stats?
AI-Assisted PRs % indicates the proportion of pull requests that involved AI coding tools, providing a direct measure of AI's impact on the code produced that week.
How does the AI Usage split categorize engineers?
The AI Usage split categorizes engineers into Heavy, Medium, or NA users based on their level of activity with AI tools.
Why is AI-Assisted PRs % considered more meaningful than WAU?
AI-Assisted PRs % is more meaningful than WAU because it reflects the actual integration of AI in coding output rather than just login activity.
How can changes in the NA count affect engineering strategy?
An increase in the NA count signals potential issues like disengagement or external factors affecting development, prompting a need for investigation.