Engineering teams in 2026 face a measurement paradox. DORA metrics expanded from 4 to 20+ including AI-specific KPIs, yet most organizations can’t answer basic productivity questions. Ninety percent of developers use AI tools and 80% believe they boost productivity, but 30% don’t trust AI-generated code. The 2025 DORA Report shows AI amplifies what’s there: strong teams improve, weak teams expose problems. The gap isn’t in available metrics—it’s in organizational capability to capture them.
The Metric Explosion Nobody Asked For
DORA’s original framework was elegant: deployment frequency, lead time, change failure rate, mean time to recovery. Four metrics that mattered. Then 2025 happened.
The 2025 DORA Report expanded to 20+ metrics: delivery (cycle time, deploy frequency), predictability (change failure rate, rework rate), investment (time allocation), and AI metrics (GenAI tool ROI, AI coding productivity). The framework went from focused lens to Swiss Army knife.
The problem? Most teams can’t capture even the basics. A Hacker News thread this week—”The Economics of Software Teams: Why Most Engineering Orgs Are Flying Blind”—hit 164 points. The consensus: organizations drown in metric options while starving for visibility. Industry research says focus on 4-5 core metrics, not everything. But few have systems to capture even that.
AI Doesn’t Fix Teams, It Amplifies Them
The 2025 DORA Report surveyed nearly 5,000 tech professionals: “AI doesn’t fix a team; it amplifies what’s already there.” Ninety percent use AI at work. Over 80% believe it increased productivity. But 30% don’t trust AI-generated code.
AI adoption improves throughput (deployment frequency, lead time) but increases stability problems (change failure rate, rework). Why? The verification bottleneck. AI writes code fast, but testing and code review haven’t kept pace. Developers ship larger pull requests, reviews slow down, quality drops.
DORA identified seven AI capabilities that determine success: strong version control, small batches, quality platforms, AI-accessible data, user-centric focus. Teams with these foundations win. Teams without? AI exposes the cracks. Tightly coupled architectures get slower. Weak testing creates more bugs.
The Time Sink Leaders Don’t See
Teams spend 40-50% of time on maintenance and unplanned work. Leaders think it’s 20-30%. This gap explains why velocity feels slow despite hiring more developers.
Investment balance tracks time on new features versus improvements versus maintenance versus firefighting. High-performing teams track where time goes, not just output. The answer is often brutal: half the capacity vanishes into work leaders don’t plan for.
The AI ROI Nobody Can Prove
GitHub claims Copilot delivers 55% productivity gains in certain tasks. But over half of finance executives can’t demonstrate AI ROI, and 42% of companies abandoned most AI projects in 2025.
The gap? Baseline metrics. You can’t prove gains without knowing your starting point. Most teams don’t track cycle time or defect rates before adopting AI. When leadership asks “is this worth $20/month per developer?” the answer is a shrug.
The measurement framework exists: track DORA throughput metrics before and after AI adoption, monitor code quality, identify power users, connect usage to production outcomes. But it requires discipline most organizations lack.
Product teams following best practices report 55% median ROI on GenAI. The keyword: best practices. That means strong testing, version control, feedback loops—foundations most teams skipped.
Start Simple, Add Complexity Later
Don’t track 20 metrics on day one. Start with DORA’s original four: deployment frequency, lead time, change failure rate, mean time to recovery. Add investment balance to see where time goes. Then layer in AI metrics when you have baselines.
Set up a review cadence: metrics to insights to actions. Every KPI should tie to a business outcome. If you can’t explain how a metric informs decisions, stop tracking it.
Teams with strong developer experience (DevEx) perform 4-5x better on speed, quality, and engagement, according to research on 40,000+ developers. DevEx measurement is straightforward: quarterly surveys on satisfaction and ease of development, broken down by team.
The 2025 DORA Report states: “The greatest return comes not from the AI tools themselves, but from a strategic focus on the quality of internal platforms, the clarity of workflows, and the alignment of teams.” Fix foundations. Measure what matters. Then add AI and watch it amplify strengths instead of weaknesses.
