MARCH 2026
A focused summary of AI adoption, productivity lift, and code quality for the duke.edu engineering team.
See how AI-active teams rank this week on the Exceeds Leaderboards.
The duke.edu engineering team reports 15.9% AI adoption, 0.22× productivity lift, and 3.8% code quality across recent work.
These metrics track how AI integrates into delivery pipelines, how throughput changes when assistance is used, and the health of AI-supported code review outcomes.
We analyze commits and diffs to estimate AI adoption, productivity lift, and code quality for your engineering organization.
Use these signals to understand how AI assistance fits into day-to-day development, where enablement efforts drive throughput, and how review practices keep quality steady.
AI Adoption Rate
15.9%
AI assistance is present in 15.9% of recent commits for duke.edu.
AI Productivity Lift
0.22×
AI-enabled workflows deliver an estimated 0% lift in throughput.
AI Code Quality
3.8%
Review insights show 3.8% overall code health on AI-supported changes.
The duke.edu engineering team reports 15.9% AI adoption, translating into 0.22× productivity lift while sustaining 3.8% code quality. These outcomes suggest AI-supported reviews are embedded in day-to-day delivery without trading off reliability.
Real questions engineering leaders ask about AI productivity, with live benchmarks and company-specific data.
What's a good company AI adoption rate?
duke.edu is at 15.9%. This is 27.8pp below the community median (43.7%)..
15.9%
Launch guided prompts, pairing sessions, and opt-in experiments to build confidence before scaling automation.
Does AI actually make developers faster?
duke.edu operates at 0.22×. This is 0.91× below the community median (1.13×)..
0.22×
Pilot AI-assisted grooming, ticket triage, or incident retros to create visible productivity wins.
How does AI affect code quality?
duke.edu holds AI-assisted quality at 3.8%. This is 19.4pp below the community median (23.2%)..
3.8%
Add structured AI code review rubrics and require human sign-off for critical surfaces.
How evenly is AI use distributed across our team?
AI impact is concentrated—82.5% of AI commits come from a few experts, raising enablement risk.
82.5%
Run prompt-sharing sessions, codify AI review checklists, and incentivize broad participation.
How can I prove AI ROI to executives?
To prove ROI, duke.edu needs steadier adoption, measurable lift, and consistent quality. The ingredients are forming but not yet executive-grade.
—
Start with a lighthouse project, measure cycle improvements end-to-end, and harden quality guardrails.
Unlock personalized insights across all your repositories, teams, and contributors.
Securely connect Exceeds with your codebase to get commit-level insights on AI adoption and performance.
See how top engineering organizations compare across AI adoption, productivity lift, and code quality.
% of commits with AI assistance
?
.gns.cri.nz
(20.0%)
h-its.org
(20.0%)
draad.nl
(-99585.7%)
wgu.edu
(-49562.0%)
Top 25% of teams adopt AI in 65-75% of their commits.
Cycle-time improvement vs baseline
?
rockstarwizard.ninja
(1.00×)
.ieselrincon.es
(1.00×)
draad.nl
(-9.59×)
wgu.edu
(-0.41×)
Top performers sustain 1.5× cycle-time improvements over six months when embedding AI into workflows.
Post-merge defect rate
?
gzgz.dev
(20.0%)
gwu.edu
(20.0%)
draad.nl
(-82634.9%)
inria.fr
(-2424.6%)
Top 25% maintain quality above 92% while expanding AI usage, pairing automation with rigorous guardrails.
Rankings based on aggregated Exceeds AI dataset of 1.2M commits across open-source and enterprise engineering teams (Q4 2025).
Top contributors
Top contributors combine high AI adoption and quality output. Encourage internal sharing of best practices.
Contributor
Commits
AI Usage
Productivity Lift
Code Quality
Sean Aery
Hakenmueller
Sara Gannon
mgs72
Marie
Encourage knowledge transfer from top AI users to others through internal mentoring or recorded "AI coding walkthroughs." Balanced adoption across the team typically improves overall performance by 12-15%.
Faster delivery
1.4x lift → predictable roadmaps
Safer velocity
93% quality → lower rollback risk
Equitable gains
AI less dependency on heroes
Governance
Depth monitoring audit-ready
Exceeds AITurns these insights into daily coaching and automatic alerts, helping managers balance speed with sustainability.