Central teams optimise well and get ignored; federated teams own outcomes and drown in tooling. The mature answer is a hub-and-spoke split by decision type, and the interview tests whether you can say exactly who decides what.
an account raises the per-topic limit · no card