create index concurrently lower( jsonb->>'name' ), drop old_indexconcurrently, ensure that where/on/group by conditions use lower(jsonb->>'name' ), then take lunch.
What's your concern with this process?
>