Most engineering teams are reporting faster cycles since adopting AI coding tools. Most engineering teams are also working without a way to verify whether those gains are real, sustained, or showing up. Today, Navigara, the engineering performance platform, is introducing Engineering Throughput Value (ETV), a per-commit metric built to test AI-era engineering claims against the codebase itself. ETV is the metric behind Navigara�s commit-level analysis across Microsoft, Google, Cloudflare, OpenAI, Meta, and Vercel�s OSS development.

The widening measurement gap
The industry�s numbers tell a story of disagreement, and the disagreement is bigger than the headlines suggest. GitHub reports 55% faster task completion with Copilot in controlled trials. Gartner projects 25 to 30% productivity gains by 2028 for teams that apply AI across the full software development lifecycle, up from the 10% Gartner attributed to code-generation tools in 2024. The "26%" figure quoted in trade press coverage falls within Gartner�s 2028 forecast range of 25 to 30%, not a current measurement. The current Gartner figure for code-generation tools is 10%. A separate 2024 Gartner survey of 724 respondents found that only 34% of teams using generative AI reported high productivity gains.

Three numbers, all cited as if they describe the same thing. One is a controlled trial measuring isolated task speed. One is a forecast for three years from now. One is what teams are reporting today. The picture gets harder from there. The 2025 DORA report found that AI adoption correlates with higher throughput and lower delivery stability. A study by the nonprofit METR found that experienced open-source engineers took 19% longer to finish tasks using LLMs: different scopes, different cohorts, different time horizons, different answers. Finance teams are being asked to budget against one of them.

"DORA metrics are old-school and outdated methods to measure engineering performance,� said Jirka Bachel, founder of Navigara and a former CTO. �As we enter the AI transformation in engineering, we also need to change the way we think about measurement and productivity.�

Where ETV fits
In February, OpenAI CEO Sam Altman, speaking at the India AI Impact Summit, called out the practice of companies blaming AI for layoffs they would have made anyway. �There�s some AI washing where people are blaming AI for layoffs that they would otherwise do, and then there�s some real displacement by AI of different kinds of jobs,� Altman told CNBC-TV18 at the summit. The pattern is measurable. Of the 108,435 US job cuts recorded in January 2026, Challenger, Gray & Christmas reports that AI was explicitly cited in 7,624 of them, or 7% of total cuts for the month. Altman named the pattern as it appears in workforce decisions. The same pattern holds in engineering, where AI is now credited with productivity gains that cannot be traced to a specific workflow change, a control mechanism, or a measurable outcome for the team using it. Navigara built ETV to test those claims directly, scoring each team�s commits the same way every quarter against that team�s Pre-AI Baseline. Whatever AI did or did not do for the team appears in its own ETV, against its own history, an aim to add a number to a conversation that currently runs on opinion.

What ETV measures
ETV is a unit of performance derived by a measurement engine that reads code the way a senior engineer does: not just what changed, but what it meant, where it landed, and whether it touched the architecture. The Engineering Throughput Value applied per file, per merged commit, combines five factors:

�For the first time, we can measure business outcomes directly from code,� said Viktor Stiskala, CTO at GTO Wizard. �Just connecting Navigara to JIRA gave us the context and clarity we needed.�

Three things ETV verifies:
Value over activity: ETV measures what was shipped and what kind of work it was. It doesn�t count tokens, lines of code, or pull requests. Anchored to the team, ETV sets the baseline based on the team�s own history, the year before AI was adopted. The reference point is the team itself, not an industry average nor a vendor-supplied study. Code history sets the baseline for every team. Commercial customers can set the date of their pre-AI Baseline cutoff, anchored to the date their team adopted AI tools. Built for the room outside engineering: ETV translates engineering work into Growth (New value), Maintenance, and Fixes language that a CFO, board, or CEO can act on without reading the codebase.

The methodology is published at https://research.navigara.com/methodology, where the scoring engine, sub-score definitions, and decay factor calculations are available for review.

About Navigara
Navigara is the performance platform for engineering managers and engineering leaders. The company built ETV (Engineering Throughput Value) to give engineering organizations a single number they can stand behind, in language the rest of the business can read. Headquartered in San Francisco with operations in Europe. More at navigara.com.

Media Contact
Michal Habdank-Kolaczkowski
michal@navigara.com