OpenAI CFO unveils AI scorecard for measuring business ROI

Sarah Friar, Chief Financial Officer of OpenAI, has introduced a practical AI scorecard intended to help organizations measure the return on investment from their artificial intelligence initiatives. The scorecard is built around four core metrics designed to capture different dimensions of AI business value. The first metric is useful work — assessing the actual quality and relevance of the output the AI system produces. The second is cost per successful task — measuring the financial efficiency of completing a given objective. The third dimension is dependability, which tracks the reliability and consistency of the AI system's performance. The fourth metric is return on compute, evaluating how much productive output is generated relative to the computational resources consumed. Together, these four measures provide a comprehensive framework for organizations to evaluate whether their AI investments are delivering measurable business impact, moving beyond early experimentation toward data-driven deployment decisions.

Measuring the financial impact of AI has become a pressing concern for enterprises. Companies investing heavily in generative AI tools and integrations often struggle to quantify whether those investments translate into concrete business value. By introducing a structured scorecard with four measurable dimensions, OpenAI is addressing a real gap in how organizations evaluate AI ROI. The framework moves beyond abstract claims about efficiency and focuses on tangible outcomes: whether tasks are actually being completed successfully, at what cost, whether the system performs reliably, and what productive output is generated per unit of computational resource spent. This kind of practical guidance helps CFOs and business leaders move beyond pilot projects to real deployment decisions.