we are currently evaluating how we could ground the visitors on the surface. In one of our older games we used Unity NavMesh and there we could just ground the Agents on the navmesh itself, including y positioning. We are not able to find a similar way in A* - any hints on this?
We’d like to avoid placing Physics everywhere (performance, workflow) since it is not required for something else.
Have at a look at the AI examples. They use a raycast to ground the agents. If you do not want to use raycasts (which should not impact performance) then you will need another way of querying the height of the ground at a given position. This does not need to be a runtime thing, you could calculate it upfront (which is effectively what the navmesh will be doing to build its height mesh).