■ SAFE RISK ■ Manufacturing & Production
Sensors already run the fermentation curve better than a clipboard ever did. The toji's role survives because the decisions that define a brewery's house style are aesthetic and reputational, not thermodynamic.
“AI monitors fermentation. But the toji's intuition has centuries of tradition.”
Our AI replacement risk score — how we score jobs
A toji runs the entire brewing season: rice polishing ratios, washing and steeping timed to the second, steaming, then koji production in a hot, humid room over roughly two days where mould propagation on rice is checked by hand and nose at intervals through the night. Then the shubo starter, the three-stage main mash addition, temperature management across weeks of fermentation, deciding the pressing moment, and blending across tanks. The toji also manages a seasonal crew, negotiates rice contracts and carries responsibility for a product the brewery's name rests on.
Instrumentation has advanced fast. Continuous temperature, gravity and alcohol monitoring with automated cooling, machine koji-making equipment, automated rice washing and steaming lines, and predictive models that forecast fermentation trajectory from early-stage data — many mid-size breweries already run these. Some producers have publicised AI-guided brewing programmes that recommend temperature adjustments. Yeast selection, quality-control analytics and inventory forecasting are straightforwardly computational. For consistent, high-volume production, automation improves reliability and reduces the number of overnight shifts that made the job brutal.
The judgement that remains is about intent. Deciding this year's rice needs a different steeping given the harvest, choosing to push a mash toward a riskier aromatic profile, tasting mid-fermentation and deciding to press early — these define what the brewery is, and there is no objective target to optimise against. Koji making by hand remains central at premium tiers precisely because the variation is the point. Add heritage, apprenticeship structures and the shrinking domestic market and you get a role that is more instrumented but still human-led. Our risk score of 21 reflects automation of labour, not of authorship.
Automatability: our editorial assessment of current and near-term AI capability
Automation is already standard in larger breweries and spreading. By 2030 expect sensor-driven fermentation and machine koji to be normal even at mid-size producers, reducing crew sizes substantially. The toji role itself persists past 2040 at premium and craft breweries where style is the product. The industry's bigger pressure is declining domestic sake consumption, which shrinks the number of breweries regardless of technology.
It can run a fermentation to a defined target reliably, and several breweries already use predictive models for temperature management. What it cannot do is decide what the sake should be — the aroma profile, the acidity, the risk you take with a particular rice this harvest. Those are house-style decisions with no objective optimum.
It is shrinking, but automation is not the main cause. Seasonal toji guild systems have been declining for decades as breweries hired year-round staff, and domestic sake consumption has fallen sharply. Technology actually helps here by making brewing less punishing physically, which widens who can enter the trade.
Rice washing, steeping and steaming lines, continuous fermentation monitoring with automated cooling, machine koji equipment, and analytical quality control. These are common in mid-size and larger breweries. Hand koji, pressing decisions and blending remain human at premium tiers because that is where the differentiated product comes from.
If you want it, yes, but go in understanding the industry economics. Domestic demand is declining while export and premium categories grow, so breweries with a clear identity and international reach are the safer employers. Learn both the traditional craft and the instrumentation — brewers who can do both are the ones being hired.