“Division of labor and data collection”

I've often been told Dassai is made by machines. The truth is that we actually employ much more staff than other producers. I guess that if people think we are industrialized, because of how much our company is structured. Because we do not rely on individuals, but rather on how effective they should be working together. Most producers begin their sake-crafting process in the winter, starting at 5 or 6 o’clock in the morning, every day without resting. People would picture the white breath of the sake making workers in the morning, and by picturing this kind of scene – maybe feel that there is some sort of art in it. The workers would work altogether on the same task: first washing the rice, then steam it before sprinkling koji spores on it. Because these processes are done by the workers altogether, it might look like some sort of traditional village gathering to a foreign eye. It probably looks very authentic, but I really think that all of this is just a waste of time.

Things are very different in our case: I decided to divide the labor between workers since 1999 when we started making sake by ourselves. It has been more efficient and in turn, helped the workers to get more technical at what they were asked to work on. For example, we get better at washing rice every single day. I also decided to work just like any company: I wanted my staff to craft sake from 8:30 a.m. to 5:30 p.m., to have them get a workday like any other employee of any other company. The fact is that, if you divide the amount of work needed for a day, nobody has to start 5 or 6 a.m. to craft sake.

I was once told by a young worker that he could put up with being busy, but that he couldn't stand being forced to stay around when there was no work left to be done. It's normal for workers to come on Sundays as well, even if there is not much work to be done. Some of the workers would start drinking sake from noon, some would not want to have to put up with it, and leave. I decided that in order to hire skilled workers that would stay with us, I had to give them at least 105 days off per year. So I first decided to give everybody the Sundays off, so that they could at least rest every week. I believe that sake making should not require people to work more than necessary: I wanted to make sake rationally and efficiently. And when we tried it, it actually worked – and worked better.

I also started to collect thorough data concerning the sake-crafting process. Back in the days, we were making decisions about whether to continue the fermentation, hold it back, or stop it based on our intuition, on a hunch. The problem was that when the sake would turn out to have a bad taste, the toji and his workers would just give a bunch of excuses, such as that it was because the rice or the weather was not good this specific year. Anyway, because sake was made on a hunch, its quality could not possibly be stable. Of course, there may be an incredibly skilled toji out there but in our case, we are just ordinary people. We had no choice but to collect data on the process to analyze it. We started to collect data on rice solubility, glucose concentration, etc. The data is collected, stored, and shared among employees. Once the data is compiled, we can understand what is the situation of the fermenting sake to date. Even so, we can never be sure of how the fermentation will react the following day: then it is a hunch, but a hunch based on data.

Right now, we have begun experiments with Fujitsu to see if artificial intelligence can be used to make decisions on the fermentation. If it is possible, it would also be thanks to the accumulation of data. However, I don't think that AI is advanced enough to be able to make judgments on what might be the most subjective judgment: “is it delicious?"

Next story: "Finding a distinctive aroma"