Wand AI CEO Proposes Agent-Hour to Measure AI Labor
Rotem Alaluf proposes the agent-hour and agentic GDP to track AI productivity as a workforce rather than a software expense.
Rotem Alaluf, CEO of Wand AI, argues that companies are failing to manage AI agents as a true workforce by treating them as software expenses. Alaluf asserts that while CFOs can track model fees and cloud consumption, they lack visibility into the actual business outcomes and productivity of AI agents.
To address this gap, Alaluf proposes the agent-hour as a basic unit of AI labor. This unit represents a measured hour of an agent operating under defined governance, which should be tracked alongside human supervision and verified output. Alaluf warns against allowing agents to certify their own success, stating that "A workforce you cannot measure is a workforce you cannot manage."
Alaluf suggests that AI value should be measured by the work it enables that was previously too costly to perform, rather than solely by the headcount it replaces. On a broader economic scale, Alaluf introduces the concept of agentic GDP to analyze the economic value attributable to agentic labor within national accounts.