Mapping the Skill Architecture of Generative AI Occupations: Evidence From Global Job Postings

Generative artificial intelligence is creating new occupational fields in which working with these systems is the core task rather than an occasional tool, yet most labour-market discussion remains theoretical and disconnected from concrete skill requirements. This study addresses this gap by examining how employers structure skill demands and roles in GenAI-intensive occupations in mature labour markets. Using 703 job postings from North America, the European Economic Area and the United Kingdom, postings were coded by business use case, technical and governance-related skills, and occupation based on the International Standard Classification of Occupations and analysed through skill co-occurrence patterns. The results reveal a clear competency hierarchy, with cloud infrastructure, model deployment and security and compliance forming a shared backbone across use cases, while frontend and prompt-interaction skills remain niche. Employers typically demand recurring skill bundles rather than isolated skills, anchoring technical core roles, while business-facing roles concentrate on governance and decision-support applications. Overall, these labour markets are consolidating around common infrastructural bundles while simultaneously producing specialised organisational niches, offering practical insights for curriculum design, workforce development and organisational planning.