I recently read that Subway in Japan has been introducing touchscreen self-ordering kiosks.
At first, I was surprised.
Subway, to me, has always been one of the clearest examples of a conversational ordering experience. You choose the bread, vegetables, sauce, and toppings while the staff prepares your sandwich in front of you. The interaction is part of the product.
But the more I thought about it, the more it made sense.
Subway's core value is customization. You can make the sandwich your own. But customization only works when users feel comfortable choosing.
In a face-to-face ordering flow, the options arrive one by one, in real time.
What bread would you like? Would you like it toasted? What vegetables? What sauce? Anything else?
For someone who already knows the flow, this can be fast and easy.
But for someone who is new, unsure, or simply wants to think for a few seconds, the same flow can become stressful. The staff is waiting. Other customers may be waiting behind them. The user may not know what all the options are. They may want to ask a question, but feel that doing so would slow everyone down.
This is especially important in Japan.
In many Japanese service contexts, users are highly aware of not holding up the line. Even when nobody says anything, the pressure is there. A user who needs time to choose may feel exposed. The problem is not the number of options. The problem is having to make those choices in public, in sequence, under time pressure.
A touchscreen changes that.
It makes the options visible. It lets users compare. It lets them go back. It lets them choose without speaking. It gives them permission to take a little time.
In that sense, the kiosk is not just a cost-saving tool. It is a UX layer around customization.
Self-service is often framed as less human. But in some contexts, it can make the experience feel safer. It removes the social pressure around the decision, while preserving the user's ability to customize.
The important point is not whether human service or self-service is better.
The important point is matching the interaction model to the pressure of the task.
For Subway in Japan, touchscreen ordering may work not because users want less service, but because they want a more comfortable way to choose.
Customization is not a feature until users feel safe using it.