Sarah Friar, CFO of OpenAI, has introduced a practical scorecard intended to help measure the return on investment of AI. The framework centers on four dimensions: useful work, cost per successful task, dependability, and return on compute.
Why it matters
By proposing standard measures such as cost per successful task and return on compute, the approach offers organizations a structured way to assess whether AI investments deliver measurable value rather than relying on general impressions.
Who should care
Business leaders, finance teams, and others responsible for evaluating AI spending may find the scorecard relevant to their efforts to quantify outcomes and dependability.