Not to be myopic, but we're pretty much "there" today with modern phones and the underlying infrastructure.
I mean, can you do Dick Tracy video calls on a smart watch yet? I think so, but I'm not sure.
Voice recognition is still, frankly, pretty bad. You can't just "talk to it", it's still fiddly, still formal. I have to "Hey Siri, put pancake mix on my groceries list." "ok, I added pancake mix to your groceries list." "Oh, and apples." ".....". "Hey Siri, add apples and oranges to my groceries list" "Ok, I've added 'apples and oranges' (a single entry) to your groceries list". "Hey Siri, add....uhhl..." "boop, I'm sorry I dont' understand you meat sack!" "Right, and it's MY FAULT you silicon dit."
Once we get in to Augmented Reality, the game changes, that's a new frontier. But beyond adding real voice UNDERSTANDING, hard to imagine what's missing from the modern mobile devices.