New Index Reveals Actual AI Delegation Patterns in Occupations

This research introduces a new metric, the Agentic Adoption Index (AAI), to quantify the actual delegation of tasks to AI by workers, rather than just the potential for AI exposure. The AAI is derived by analyzing the semantic similarity between approximately 53,000 agent skill specifications from the Manus Skills Marketplace and 18,000 O*NET task statements, aggregated to the occupation level. The study reveals three key findings. First, the occupations showing high delegated exposure to AI are distinctly different from those previously identified as "at risk" by pre-AI frameworks, indicating a divergence between potential and actual AI integration. Second, the AAI more closely tracks what AI *could* do rather than what workers *currently* use it for, suggesting a gap between capability and adoption. Third, the AAI shows a peak below the top of the wage distribution and at the bachelor's degree level, declining at both extremes. While technical availability explains much of this variation, it does not account for the shortfall among the most educated occupations. This suggests that factors like work resisting advance specification or professional discretion over codification play a significant role in who adopts AI, highlighting that feasibility alone does not drive adoption.