The Next-Gen Drive-Thru Is About AI, Not Mobile Ordering
thespoon.tech
thespoon.tech
The two important points made in the article are that fast food places have issues with timing mobile orders, and many store managers notice that mobile pickup facilities are overprovisioned relative to their usage.
The first is a solved problem in the world of takeout pizzas; in fact, such pizza places often have minimal storefronts and waiting space. But pizza keeps well in a box under its own heat, and heatlamps have become unfashionable in fast food as everyone wants to move upmarket and made-to-order. But there's also a UX solution: detect or ask the customer via the app when they're ready for the order to be made.
The second point is ironic: all these companies pushed into mobile ordering as a top-down directive thinking that's what users want. They'll push into AI or whatever hype next the same way. It's bullshit. People want low prices (McDonald's, Taco Bell), or they want the appearance of quality (Panera). Blowing a ton of money on some half-baked ML ordering system is a waste, when they've already cut their cashiers by putting up slow kiosks and put all their ordering UI on everyone's phone. But the people selling this to clueless executives spinning their wheels stand to make a decent chunk of change until they cash out to some equally-clueless established point-of-sale company.
Let's take McDonald's as an example. I've worked there as a crew member and, many ironic decades later, currently assist some of their OEMs in the kitchen equipment area.
McD has always been a JIT-based manufacturing/assembly company. And their production window isn't a week or a day: it's 10 minutes. Back in the old days when you walked in and there was a bin full of burgers, those were all made to be sold or tossed after 10 minutes. How do you manage that without wasting a single burger?
Managers would keep insanely detailed logs of sales by the hour or even half-hour to try and assemble some kind of historical record that could be used to anticipate future demand. They would even record the traffic level outside and weather conditions. It worked pretty well, but one rogue schoolbus full of hungry teens would blow that forecast out the window.
The newer systems try to smooth some of that out with heated holding cabinets and faster toasters and et cetera, but there are some things you really can't batch and hold for very long, like fries. Customer demand has also required some of the batched systems go away, like the Quarter Pounders being grilled as they are ordered.
So when a batch of fries takes 5 minutes to cook, and you know there's a wave of people arriving, any technique to help anticipate that and trigger a production cycle sooner is ideal. It benefits producer and consumer alike.
Mobile ordering isn't a panacea, and in fact the current system isn't very usable at all. But they're trying. And I do know second-hand that they're aware of geofencing but there are complications to a system like that as well.
This AI BS is just a scam designed to dupe clueless executives out of their money. It’s just window dressing. The real problem? Constant delays in the drive thru. People can order just fine already, what pisses them off is sitting in the drive thru for 10 minutes waiting for “fast food”. Not to mention the price increases and a lack of new novel menu items.
So the temptation for fast food sellers is to make customers order ahead. This is an economy for the store, not a convenience for the customer. Thus, efforts to encourage or coerce customers to order online.
This leads to a proliferation of ordering apps, all different. That will be followed by consolidation: "Uber Eats", "Amazon Fresh", "Google Chow".
They say mobile ordering is bad, so the solution is AI ordering, and a touch screen.
Is this just a journalistic misinterpretation of neural networks?