I remember when the drive through audio quality was often pretty bad. I think it was in the 80s when they adopted audio systems with radio, and the order takers could move around as needed. A Ham friend at the time found the frequency they used and placed his order well before going into the drive through.
(I had to remove the "cheeseburger and large orange drink" sketch because the fucking bad slashdot system claimed "lameness filter encountered".)
Now, there are stories about the order audio going off into different cities and the orders handled in something like call centers. I guess this means one less training skill for the local workers, and they can focus more on getting the food ready to serve. Using computers to interact vocally with the customers seems like the next step. But if the local fast food joint was getting reliable order info from real people somewhere else, and now have to deal with order mistakes from a computer, that is going to throw gravel into their well oiled machine.
Looking at this from the customer side, I can see my order on the screen as I say what I want. So any mistakes should have an easy way to object and have the computer delete the bad entry and try again. This would be annoying, after having decades of better service with real people. The top level execs probably think the customers will get used to the new system and just accept it.