Hi, one basic way to add more variance would be to let your AI not all move the the same location (the target actor), but sample a location in a random radius around the target and let your AI move to that location instead or to go even further from that, maybe use an EQS query to sample a location around the target and let it move to that (but I would start with a random location to see that it works first).
Of course then you need to check on a timer, that the location the AI is currently moving to is still close enough (whatever “close enough” would be for you) and if not, then run the logic again to find a new location (so basically manually do there what MoveTo actor does under the hood).
You could even think about first sampling a location in a larger radius around the target and once the AI reached that location, let it move to a closer one.
By the way, if what you store in your blackboard is of type actor, then you can set that in the blackboard and then directly access it via GetBlackboardValueAsActor. There is then no need to get it as object and casting it to actor.