A central effect of automation is to transform jobs - shifting their task content. We develop a general-equilibrium model of this process. Occupations bundle tasks; workers possess task-specific skills and sort by comparative advantage. When a task is automated, remaining tasks gain in importance, so wage effects depend on workers’ full skill profiles. We estimate the distribution of task-specific skills and project individual-level wage effects of generative AI automation. Moderate exposure benefits workers on average but high exposure harms them, with large dispersion within occupations; the return to social skills rises, that to analytical skills falls; and low-earners gain more than high-earners. Job transformation drives these results.