How will AI redraw the boundaries between specialist roles?

2026-10-16 · question

AI automates tasks, but organisations are structured around roles: product, architecture, engineering, QA and the handoffs between them.
If half of two specialists' tasks are automated, the remaining halves might form a new specialism rather than eliminate specialisation.
As people acquire adjacent capabilities, they can take responsibility for larger units of work, including more of the customer-to-feedback loop. Context switching and depth still limit this.
What looks like generalisation within the old division of labour may actually be specialisation around a larger unit of work.
What determines the boundaries of these new specialisms?