TL;DR

• Coding ran 4.68 days inside an 11.7-week delivery cycle at one enterprise health tech client, about 6% of the calendar.
• AI budgets, the seat licenses and the model bills, land almost entirely on that 6%.
• The cycle is set by the other 94%: intake, estimation, handoffs, review, QA, and deployment, most of it still manual.
• Teams that compress the 94% build two things the tools don’t supply: capability and conduct.

Where the weeks actually go

We measured the delivery pipeline of an enterprise health tech client. The active coding stage ran 4.68 days inside an 11.7-week delivery cycle, about 6% of the calendar. AI has been aimed almost entirely at that 6%. The other 94%, intake, estimation, handoffs, code review, QA, and deployment, is where the weeks actually go, and most of it is still manual.

The money follows the same pattern. The AI budget lands almost entirely on the coding stage, those same 4.68 days. A team can put its whole AI investment into that window and the 11.7-week cycle barely moves, because the cycle is set by the 94% the budget never touches.

One caveat before the numbers that follow: the lifecycle figures are our own analysis of a single client’s pipeline. The rest is a field reading built on public research plus our work with regulated teams, not a controlled benchmark. We’d rather show you the shape of the problem honestly than dress a client anecdote up as a study.

Usage went up. Outcomes didn’t follow.

Almost every team now uses AI; almost none can say whether it’s making them better. About 90% of technology professionals now use AI day to day, per DORA’s 2025 State of AI-Assisted Software Development report. Yet the same research program found a 25% rise in AI adoption associated with a 1.5% drop in delivery throughput and a 7.2% drop in stability.

That result stops looking strange the moment you hold it against the 6% number. AI made the typing faster. The cycle is not made of typing.

“The cycle time you feel is end to end. Doubling the speed of 6% of the timeline changes almost nothing; the real gains come from cutting wait and handoffs across the other 94%.”

Sapan, co-founder, Incubyte. From the AI-Fluent SDLC brief.

What it takes to reach the other 94%

Reaching past the coding stage is not a tooling question. It’s a depth question: how much of a single unit of work can AI carry on its own? In our fluency assessment we read this as five levels, from prompting to autonomous across the cycle.

The jump from level 3 to level 4 is the one that matters. That’s where AI stops working inside the coding stage and starts reaching into the 94%, pulling from the ticket, settling acceptance criteria, and moving work toward deployment.

Most teams stall before that jump. The higher levels are available; what’s missing is the ability to review and govern fast enough to trust an agent that far. Coding-agent adoption still sits between 15.85% and 22.60% across 129,134 GitHub projects, even as about 80% of new developers reach for Copilot in their first week.

Fluency, not adoption

What turns adoption into outcome is capability and conduct. Capability is what your people can do with AI across the whole lifecycle: the judgment to direct and check what AI produces, well beyond the coding stage. Conduct is what your systems hold your people and your agents to while they do it: your standards, the regulations, and the data boundaries, enforced as the code is generated rather than checked after.

Capability lets your team push AI across more of the lifecycle; conduct lets them do it without the cycle blowing up in review or audit. Both can be built, and both can be measured. The 94% is reachable. It just isn’t reachable with a seat license.

The full argument, the fluency stack, and the measurement model are in our brief, The AI-Fluent SDLC. It’s written for CTOs and VPs of Engineering at regulated and health tech software organizations. Read the brief.

Or start with a mirror instead of a paper: the AI Readiness Assessment places your team on the fluency quadrant in six questions, with a full report by email. Take the assessment.

Sources

• DORA. 2025 State of AI-Assisted Software Development Report. Google Cloud, 2025.
• DORA. Accelerate State of DevOps Report 2024. Google Cloud, 2024.
• Robbes, R., Matricon, T., Degueule, T., Hora, A., & Zacchiroli, S. Agentic Much? Adoption of Coding Agents on GitHub. arXiv:2601.18341, 2026.
• GitHub. Octoverse 2025. GitHub, 2025.
• Incubyte. Enterprise Health Tech Client Delivery Pipeline Analysis (internal analysis), 2026.