Marukai, a 108-year-old miso maker, achieved 86% AI usage among 290 employees.
The company paired clear rules with training and peer support.
A 70-year-old president's active participation helped drive adoption.
What happened
Marukai, a 108-year-old miso maker in Mie Prefecture, gave all 290 employees access to Google's AI, Gemini. The login-based usage rate has reached 86%.
Why it matters
The company succeeded where many AI rollouts stall by pairing rules with training and shared know-how. A 70-year-old president joining study sessions and a manager building a reservation tool with Google Apps Script helped change workplace attitudes.
What to watch
Marukai starts with a 25-license pilot before company-wide rollout in January 2025. Managers now encourage staff to try Gemini and share effective prompts through chat tools, aiming to prevent over-reliance on AI output.
Ask the AI about this article →
Marukai's approach addresses a common corporate problem: providing AI access without seeing actual usage. The company's strategy centered on three pillars: explicit rules, shared knowledge, and education. This structure gave employees clear boundaries while encouraging experimentation.
The 86% login rate suggests the approach worked, though the article notes this measures logins, not necessarily the depth of use. The president's participation in study sessions appears significant for signaling that learning AI was a priority at all levels, including leadership.
The company's emphasis on responsible use—banning confidential data input and teaching employees to verify outputs—reflects a broader concern about AI risks in business settings. By positioning AI as a tool with both potential and limitations, Marukai aimed to build sustainable habits rather than just initial enthusiasm.
For example, today's edition would include:
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. Q&As are published on this page for other readers too.
Google launched two large language models, Gemini 3.8 Flash and Gemini 3.8 Flash Cyber
CrowdStrike has established a cyber superintelligence lab to develop security-specific frontier AI models, aim…
JPC Connectivity said AI infrastructure demand strengthened in Q2 2026, boosting orders for optical communicat…
Auras Technology's chairman, Steve Lin, said the company is expanding from GPU platforms to ASIC-related produ…
Google released the AI models Gemini 3.8 Flash and Gemini 3.8 Flash Cyber on September 2
AI's workplace efficiency gains often disappear because companies never define where saved time should go
