You are correct that a lot can be automated, but if there is anything that cannot be, humans have to be in the cockpit. And if humans have to be in the cockpit, then we actually don't want as much automation as possible. [1] See also [2].
The biggest problem I see with automating helicopter flight is that the hardest part of an autorotation is picking a spot to land. I don't think that can be automated. You would either have to have super high resolution radar pointed at the ground or high resolution computer vision that can identify flat spots which are as varied as the human race. I don't think either is feasible anytime soon. And that's without even considering the higher power draw they would require, as a sibling comment mentions.
Then you have to consider what happens in the case of full power outage, including no battery. In that case, you have effectively doomed the people onboard if there is not a real pilot.
[1]: https://www.nbcnews.com/news/us-news/airline-pilots-depend-t...
In a smaller aircraft, you're relying a lot more on your forward momentum (i.e. forward airspeed) to cushion your landing, whereas in a larger helicopter the rotor inertia has enough energy such that you can place it straight down on pretty much any flat surface the system can find. Garmin already has something called Smart Glide for fixed wing aircraft where it calculates engine out trajectories, and so I don't see this as something impossible with a helicopter. Especially coupled with a wire strike cut (basically just a sharp knife sticking out of the front of the helicopter that just cuts any cables you run into).
I'd think given the use cases for some helicopter flights (e.g. sightseeing tours) where you're typically in VFR weather there's certainly an opportunity to have automated flights. I agree we're not there yet, but I don't see it as impossible in the near future.
(I mainly fly fixed-wing, but I have about 50 hours in an R22)