Planning with learned world models can involve actions beyond the physical world, such as searching the internet, interrogating databases, running simulations, or using tools like calculators.
If the world the system evolves in is not the physical world, but is the world of let's say the internet or some sort of world of where an action consists in doing a search in a search engine or interrogating a database, or running a simulation or calling a calculator or solving a differential equation, how do you get a system to actually plan a sequence of actions to give the solution to a problem?