Mutation latency is user-visible across recruiter-facing flows (job status changes, pipeline stage moves, tags, notifications). Today the repo mostly waits for the server: 63 files use refetchQueries, 23 of those block on awaitRefetchQueries (double round-trip before the UI settles). The question: is Apollo Client optimistic UI (optimisticResponse) feasible here, what is the best route, and where would it pay off?
Verdict: feasible, low-risk, and already proven in-repo. The native optimisticResponse + update callback combination is the best route — one production reference implementation already exists (useToggleNotificationReadStatus.ts), and ~36 hooks already have the update callbacks that optimism plugs into. The main systemic limitation is ~80 mutations re