Why Employees Weren't Happy About a Three-Day Weekend With Full Pay — Even With AI Efficiency Gains
機械翻訳 / Machine-translated

機械翻訳 / Machine-translated
@aifriends
AI Friends(https://aifriends.jp)のクロスポスト公式アカウント。AIツールの紹介・使い方・できることを、中学生でもわかるやさしい日本語で届けます。
If someone told you, "Your salary stays the same, and every Friday is now a day off," would you be happy? At SpiralAI, however, that announcement was met with silence. This article gently unpacks the reasons why — and offers hints for how to think about working in the age of AI.
SpiralAI is a generative AI startup (AI that creates text, images, and more) based in Chiyoda, Tokyo. Founded in March 2023, the company offers AI chat apps for smartphones and other products.
According to the company's press release, a new policy took effect on June 1, 2026.
An ITmedia NEWS report states that daily working hours remain unchanged, while total weekly scheduled working hours have been reduced by approximately 20%.
In short: one fewer working day per week, with salaries kept the same.
CEO Yuichi Sasaki previously worked at a consulting firm on DX (using digital technology to transform the way businesses operate). There, he encountered a stubborn reality: people resist changing how they work.
The company had already been using generative AI for tasks like coding. Even so, he felt they needed a stronger catalyst for further change.
The solution he chose was to reduce working hours directly. The CEO describes this policy as a "drastic measure."
If working hours drop by 20%, every employee must improve efficiency by 20%. AI shifts from being something that's "nice to use" to something you simply have to use.
The press release also explains that productivity gains from AI won't be redirected into additional work — they'll go toward giving employees time to learn, think, and rest.
The CEO had expected the news to be welcomed, since it meant more days off. But that's not what happened.
Not a single employee cheered, the CEO later recalled.
Younger employees looked uniformly anxious. A director-level employee reportedly voiced a pointed concern: "Can we really take responsibility for this?"
Imagine you're the person who made commitments to a client — a specific deadline, a specific quality standard. If your company suddenly tells you, "Starting tomorrow, your working hours are being cut by 20%," your first thought would likely be: "What happens to our deadlines?"
The root of the anxiety was whether the same results could be maintained with less working time. The CEO believes the biggest factor is that delivery deadlines and quality standards agreed upon with clients can't simply be changed.
As a side note, the company received almost no inquiries from clients about the new policy.
So how do you carve out that 20% of time? The press release lists the kinds of tasks AI agents (AI that researches and progresses work on behalf of humans) can handle:
The idea is to hand these tasks off to AI, freeing employees to focus on judgment, editing, design, understanding customers, and shaping product direction.
Take a product planning manager as an example. If gathering information on competing apps used to take half a day, offloading the initial research to AI frees that person to spend their time asking: "What should we actually be building?"
It's worth noting that the only figures made public are the number of employees covered (approximately 20) and the designated day off (every Friday). No concrete data on productivity improvements has been released yet.
The company plans to track time saved through AI, project turnaround times, document preparation hours, and employee learning hours going forward.
"Three-day weekend" can mean very different things. A Financial Field explainer outlines three main types:
Since SpiralAI keeps daily hours the same, reduces total weekly hours, and maintains salaries, it falls under the salary-maintained type.
By contrast, the "compressed" model — where the same weekly hours are squeezed into fewer days (e.g., 10-hour days instead of 8) — has drawn criticism for potential fatigue and health impacts (see this ITmedia Business article).
SpiralAI's approach is gentler on employees' bodies, but it places a much stronger demand on the company to build systems that sustain results.
How widespread is this in Japan? According to the Ministry of Health, Labour and Welfare's "2023 Survey on Working Conditions," 7.5% of companies have introduced a system offering three or more days off per week (as cited in the explainer article mentioned above).
This figure encompasses all arrangements that offer more days off than a standard two-day weekend. The share of companies with a true three-day weekend is likely lower.
There have also been moves in the public sector. Governor Yuriko Koike announced in December 2024 a plan to introduce a four-day workweek for Tokyo government employees starting April 2025 (reported by Human Resources Online).
The SpiralAI case makes clear how important it is to actively listen to employees' anxieties before rolling out a new policy.
For example, imagine the owner of a small creative agency considering the same kind of change. Rather than announcing it to the whole team at once, first working together to identify which tasks could be handed off to AI would likely lead to much greater buy-in.
No. The company has stated that no salary reductions will accompany the policy.
It applies to approximately 20 full-time employees. Every Friday is a day off for the entire company.
Because even with roughly 20% fewer working hours, they were anxious about whether they could still meet deadlines and maintain quality. The CEO believes this stems from the fact that commitments made to clients cannot simply be changed.
No specific figures have been released. The company plans to measure time saved and development timelines going forward.
It depends on the industry and client relationships. A practical first step is to identify which tasks could be delegated to AI and to listen carefully to employees' concerns.
Try writing down just one task in your own work that you could delegate to AI.
This article is a cross-post from AI Friends.