There was a significant amount of manual piloting in the Apollo missions
The guidance gets more difficult in the terminal stages and they didn’t really trust computers to safely control the spacecraft near the surface, so their solution was to have the computer fly 95% of the way down and have the crew take over for the terminal phase.
The Apollo algorithms work fine for non-manned missions as well, but you have to vet the trajectory targets more fully in simulation and add some active retargeting scheme to avoid obstacles near the surface.
Combine the added complexity of a robotic lander with groups like intuitive that have never landed one before, and this sort of thing happens