How to Realize the True ROI of Software Development – An In-Depth Analysis of the Latest DORA Research

To prove the business value of generative AI, it is essential to go beyond simple technology adoption and build the organizational systems and culture that allow AI to function effectively.

In the early stages of AI adoption, a temporary decline in productivity known as the J-curve can occur. This is not a sign of strategic failure but a natural process caused by the following three factors.

• Learning curve: Teams need time to learn new workflows and acquire advanced prompting skills
• Verification tax: As the volume of AI-generated code surges, additional review time is required to prevent hallucinations and maintain quality
• Pipeline adaptation: Downstream processes such as testing and approval need to scale to keep pace with faster development speed

According to DORA's research, 90% of respondents use AI in their work, but actual financial outcomes vary across companies. To achieve successful ROI, organizations must clearly identify where AI contributes to cost reduction, productivity gains, and stronger security, and build a realistic financial model that accounts for both visible and hidden costs.