This was absolutely the correct answer! I put each tower/object only 10cm up from the landscape and ** every ** object was seen when it should have been seen! I figured GetActorLocation used the pivot point but I didn’t figure Perception used it. I assumed it would look for the bounds or collisions of the object using a trace.
I was able to get similar results, bypassing Perception using EQS, before you gave me the answer. Worked decently but if I’m not mistaken Perception is better to use than EQS for this? Although, now I’m going to combine Perception with EQS to get the desired behaviors I need but at least now my behavior tree doesn’t need the extra EQS procedure to return it’s “seen” actors and can run independently of each other.
Thanks again, you are awesome!