Since we're talking far-out future anyway, it's not hard to imagine a delivery drone-copter that flies your pizza to you right outside your window.
Imagine having the choice of two pizza places - both charge the same price for the pizza, and taste just as good. One uses human drivers who drop it off at your door - but expect to be tipped. The other uses robotic delivery trucks; you have to pick the food up at the street, but you don't have to tip. Which would you choose? Or rather, which would the median pizza purchaser choose?
This is analogous to how a lot of automated or "self-service" services work, such as grocery checkouts[1], or petrol stations[2]. The automated service is often inferior in every way but price...but that's usually enough.
[1]: This is a bit of a special case; in my experience the automated checkouts aren't cheaper, so I only use them when they'll clearly save me time due to shorter lines. Still, time is money, so I think it's still an example of an inferior -but-cheaper service.
[2]: I understand that concerted lobbying has resulted in laws preventing people from pumping their own gas in some US states. For the sake of this post, pretend you live somewhere where self-service stations are the norm, as the are in most of the world.
1) Pathfinding is easier as it doesn't have to concern itself with roads or obstacles for most of it's route. Just fly up to a reasonable altitude (just enough to clear buildings in the way) and all the drone has to worry about is adjusting for wind and "last-mile" (your window) navigation. Once they start hitting critical mass we'd start seeing infrastructure to keep the drones moving; homing beacons and landing pads on your window, and drone towers (either standalone or grafted onto existing skyscrapers) to refuel(/charge) drones and hoist them up to a high altitude, allowing them to rely more on gliding vs. their own power.
2) Less of a chance to run someone over. Again; since it's not in conflict with existing infrastructure, it's less likely to suddenly find a 6-year-old in front of it. The flip side is that it's going to hurt if the drone fails and drops out of the sky; this can be remedied by a fail-safe (such as a parachute that physically breaks open at a certain velocity; independent of the nav systems)
3) Commercial advantages: besides the general advantages of a drone-based delivery; a flying drone doesn't get stuck in traffic as it's skipping roads entirely. Thus, using a flying drone will give you a significant advantage over your ground-based (be they human or drone) competition re: delivery times. If they aren't faster; they're at least going to be far more predictable.