DORA

The ROI of AI-assisted Software Development, published in April 2026 by the DORA × delta team at Google Cloud, runs to 60 pages and ships with an interactive calculator at dora.dev/ai/roi/calculator under CC BY-NC-SA 4.0. It continues a line of work reaching back through the 2020 ROI of DevOps Transformation report, the 2025 State of AI-assisted Software Development, and the DORA AI Capabilities Model. Nathen Harvey leads the group.

The pivot claim is that AI is an amplifier: it magnifies the strengths of high-performing organizations and the dysfunctions of struggling ones. Buying licences accomplishes nothing on its own. Injected into a system with manual testing, bureaucracy, or fragmented data, AI accelerates the accumulation of technical debt. The measurement rule is stated flatly: "we don't measure AI by the code it writes but by the bottlenecks it clears."

The J-Curve of AI value realization names the temporary trough (learning curve, verification tax, pipeline adaptation) that precedes exponential growth, a tuition cost of transformation to be budgeted explicitly. The sample calculation, 500 FTE at $176k salary with 12.5% time saved per developer, returns $11.6M of value against $8.4M of investment: 39% ROI, payback in eight months. DORA recommends against a headcount-reduction strategy, argues for reinvesting freed productivity into innovation, and asks that the Internal Developer Platform be run as a product.

That 12.5% is a financially defensible floor, well under the ratios practitioners report. In Louis Claxton's AI-native SDLC playbook of 21 August 2026, the delivery performance DORA measures appears as a lagging indicator of the CI/CD play.