As long as a mutation isn't strongly maladaptive, it can evolve prior to its being useful.

They didn't need to come about at the same time. Photosensitive proteins (opsins) and cellular motility both predate multicellular life entirely. Even single-celled euglena detect light and swim toward it with no nervous system at all. In early multicellular animals, cells were already chemically signaling their neighbors. A photosensitive cell releasing a signaling molecule near a contractile cell isn't a coordinated miracle. It is just two pre-existing cell types sitting next to each other in tissue, which is what bodies are. Natural selection then refines that crude coupling because even a tiny, noisy light response is better than none.

Each piece, light-sensitive proteins, cell-to-cell signaling, contractile cells, evolved independently and for other reasons long before being co-opted into anything resembling vision. The question "how could A and B arise simultaneously?" dissolves once neither A nor B was new.