OpenAI CFO Proposes ROI Scorecard via Cost per Successful Task
July 17, 2026
OpenAI CFO Sarah Friar suggests measuring AI utility using specific metrics: cost per successful task, dependability, and return on compute. This framework moves beyond simple token usage to focus on actual work completed per unit of expenditure.
HOW THIS AFFECTS YOU
●
founderYou can use these metrics to justify AI spend to stakeholders.
●
investorThis provides a framework to evaluate the actual efficiency of AI-driven startups.