Well yes but you are moving the goalposts
Long range
enroute nav works pretty well with NDBs because the field distortion caused by coasts and terrain tends to be only local to the beacon, so if you are 100nm away from the beacon, it will be accurate. On my system I have found enroute NDB nav to be more or less as accurate as VOR tracking. I can see the old NDBs, radiating tens of kW, worked very well in the old days, and there were no VORs.
And anybody with the briefest training can enroute track towards an NDB. Especially when enroute when there isn't normally much to do.
But earlier on you talked about NDB
holds. These, especially if flown with the 'gates' beloved by IR examiners, are just plain hard work.
Autopilots don't come into this, though obviously they drop cockpit workload by an order of magnitude. But I think that if somebody has an AP, they probably also have a GPS....
And very few GA autopilots can fly holds. Only the relatively modern kit can. You need something like a GNS430W or 530W, driving an AP via ARINC, to fly holds - especially if you want it to compute and fly the correct pattern entry. My AP can fly anything but the KLN94 GPS doesn't contain holding patterns or even (in Europe) any curved tracks.
If I had to fly a hold (only about once in last few years) I would use the OBS mode of the GPS to give me the inbound track on the screen, and then twiddle the heading bug.