I think there are many meaningful improvement strategies which could be applied that haven't been tried yet from any administration at scale.

A lot like replacing running tech systems while they're running - it's trickier doing that than stopping and starting. But, there have been a lot of technology systems less complex than the federal gov't that have stopped, changed to the new system and restarted, only to find that the new system completely broken.

If you want to replace while running, you can set up a parallel co-running system at lower capacity, trial run samples through and compare results, then take some capacity of actual cases, basically take steps to keep growing maturity. For the govt there's actually an interesting case to do this to serve younger more tech savvy citizens, while allowing older citizens to stay on the existing system - then slowly shift capacity from old to new systems.