You need the camera location/actual HMD position instead of pawn location. You would also want a collision object attached to the camera that can interact with the AI.
You need the camera location/actual HMD position instead of pawn location. You would also want a collision object attached to the camera that can interact with the AI.