A framework for estimating AI-driven savings in software development

The rapid integration of Artificial Intelligence (AI) and Large Language Models (LLMs) into the Software Development Life Cycle (SDLC) promises significant transformations, primarily driving enhancements in efficiency, quality, and time-to-market. However, quantifying the economic return on investment (ROI) from these technologies often remains challenging due to the highly contextual and non-linear nature of software projects. This paper proposes a quantitative framework designed for software projects to estimate AI-driven effort savings. The framework systematically maps projected percentage savings derived from AI automation across major SDLC phases, incorporates project-specific contextual risk factors that modulate these savings, and utilizes elements of existing software effort estimation (SEE) models, such as COCOMO II, FP, to generate a reliable AI-adjusted effort estimate. By defining potential savings with supporting evidence and detailing critical contextual influences, this model provides a structured, data-driven approach for organizations to forecast the measurable economic impact of AI integration.