CFO Sarah Friar introduced a framework to measure AI return on investment via four metrics: useful work, cost per successful task, dependability, and return on compute. This shift moves focus from vague productivity gains to concrete unit economics. Practitioners can now quantify the actual efficiency of OpenAI deployments against specific operational costs.