AI Shaping: Turning General-Purpose AI into Productive AI Work
Kurni Kwok · Zenodo (CERN European Organization for Nuclear Research) · 2026
Recurring AI-enabled work can produce useful outputs while leaving people to reconstruct source basis, domain fit, work state and resumption logic. AI shaping is the operating-model discipline that establishes shaped intelligence to carry a reusable and iteratively improvable domain work pattern and reusable work basis under human direction. Shaped intelligence carries more of the two-sided burden — domain-practice and subject-context, and work-state and resumption — so useful output can become grounded, reviewable, repeatable and resumable productive AI work. The direct-shaping approach establishes bounded shaped intelligence without first materialising separate AI-shaping intelligence; the mediated-shaping approach uses separately reusable AI-shaping intelligence to carry the shaping pattern. This paper defines the category, supports category-fit decisions and routes readers to separate Project-managing and Product-managing evidence papers. It does not establish capability performance, measured improvement or implementation-transfer authority.