"Don't use AI for busywork."
When I first heard this, I honestly didn't understand what it meant.
Isn't letting AI handle chores the most straightforward way to use it? Drafting emails, summarizing minutes, fixing document formatting. I thought AI existed to reduce those tedious tasks.
But as you follow Takayuki Fukatsu's insights, the landscape changes completely.
CSO (Chief Strategy Officer) of note Inc., CEO of THE GUILD, CXO of Bengo4.com, and AI Strategy Advisor for Yokosuka City.
note has 12.48 million members and over 82.09 million pieces of public content. Recent semi-annual sales are 1.9 times higher than the same period three years ago. Sales per employee have increased by 2.3 times. Traffic via AI has reached approximately 4 times the predicted value from search.
What has Mr. Fukatsu been doing behind these numbers?
I'll give you the conclusion first.
Takayuki Fukatsu isn't just delegating work to AI. He is using AI to grow human potential.
Why this person is qualified to talk about AI
You might think, "Isn't he a UX designer?"
True, Mr. Fukatsu's career began in design. After working at the Flash content production company tha, he founded Art&Mobile. He then established THE GUILD and became famous for iPhone app UI design.
However, since 2022, note has shifted its management axis toward the AI field, with Mr. Fukatsu leading the charge. He joined as CXO (Chief Experience Officer) and became CSO (Chief Strategy Officer) in July 2026. The change in title itself tells the story of AI's evolving role at note.
The results are the numbers mentioned at the beginning: Sales up 1.9x, sales per employee up 2.3x, and page views up 1.4x year-over-year.
He has also authored two books: Mastering ChatGPT! The Fukatsu-style Prompt Reader (Nikkei BP, August 2024) and Meta-skills (co-authored with Kensuu and Kazuhiro Obara, NewsPicks Publishing, May 2026).
His influence extends beyond note. He is the CXO of Bengo4.com, an AI Strategy Advisor for Yokosuka City, a visiting professor at Musashino Art University, and a visiting lecturer at ZEN University.
A person who started as a designer and used AI to transform organizational metrics. I will break down his methods into seven parts.
Method ① Fukatsu-style Prompt — 4 elements that change AI output
When giving instructions to AI, do you just write "make it look good"?
The "Fukatsu-style Prompt" he released consists of just four elements:
Command (You are a [Role])
Constraints (List rules in bullet points)
Input Text (The text you want processed)
Output Format (The framework for the AI's result)
For example, when you want to create meeting minutes:
You are a corporate secretary. Based on the following constraints and input text, create concise and comprehensive meeting minutes. Constraints: Separate by headings, speakers, and key points. Organize chronologically to show the flow of discussion. Summarize each statement within 300 characters.
By writing this, the AI can simultaneously grasp "as what expert," "under what rules," and "what to process."
Since this template was published as a note article in 2023, it has been widely used by companies, local governments, and individuals. It was also turned into the book Fukatsu-style Prompt Reader (August 2024, Nikkei BP).
If you want to improve prompt accuracy, start by mastering these four elements. It drastically reduces trial and error.
Method ② Make AI your "Master" — Use it for human growth, not just tasks
This is the true meaning of "Don't use AI for busywork."
When Mr. Fukatsu brainstorms with AI, he writes something like this:
"Please teach me and help me grow by pointing out my shortcomings so that I can become a person who can create wonderful medium-to-long-term plans in the future."
He isn't asking the AI to generate a plan. He is using AI to grow into a person who can generate plans.
By giving the AI roles like "Master," "Teacher," "Yoda," or "Master Avan," it won't just give you an answer; it will point out holes in your thinking or weaknesses in how you frame questions.
To quote Mr. Fukatsu: "Many people use generative AI to execute tasks. But it's more efficient to use it to grow humans."
If you let AI write a proposal, today's proposal is finished. But if you let AI train you, every proposal you write from tomorrow onward will change.
To be honest, until I learned this, I was satisfied just memorizing the "template" of the Fukatsu-style prompt. I thought filling in the four elements was enough. But I later realized that what Mr. Fukatsu really wanted to convey was the "way of facing AI" that lies beyond the template.
Method ③ "Management should be AI" — An unbiased outside director
AI-driven customer support might save hundreds of thousands of yen a year.
One management mistake by a CEO can lose tens of millions to hundreds of millions of yen.
Mr. Fukatsu asserts that "the highest risk in small and medium-sized enterprises is actually the intuitive and arbitrary decisions made by management."
Specifically, what do you do?
You let AI participate in management meetings. Transcribe the minutes, feed them to the AI, and ask, "Are there any issues that should be discussed but are being ignored?" Have it point out decisions made based on "momentum" rather than evidence. Have the AI analyze monthly trial balances to detect early signs of cash flow deterioration or the need for price pass-throughs.
Functionally, the AI acts as an unbiased outside director.
Mr. Fukatsu distinguishes this as "Tactical-level" vs. "Strategic-level." Automatically creating greeting emails is tactical. Improving the quality of management decision-making is strategic. Most companies are working on the tactical level.
However, the return on investment is orders of magnitude higher if you stop a CEO from impulsively building a strange office building or moving forward with an acquisition without sufficient due diligence.
Method ④ Google Calendar-linked Secretary AI — Get the latest info 15 minutes before a meeting
This is a system Mr. Fukatsu actually runs.
It automatically detects company names from appointments registered in Google Calendar and, 15 minutes before the meeting, automatically inserts a report summarizing that company's latest news, press releases, and recent movements.
He uses tools like Dify, Zapier, Python, and Google Apps Script depending on the situation.
The time spent searching "did anything happen with that company recently?" becomes zero. Because you enter the meeting with the latest situation in mind, the quality of the conversation changes.
When you apply Method ③'s "Strategic AI" to daily life, it looks like this. What it's doing is mundane. But if you have five meetings a day, the preparation for all five is finished automatically. The more mundane something is, the bigger the difference when it accumulates.
Method ⑤ Meta-skills — Train human "higher-level skills" before AI
Meta-skills (NewsPicks Publishing) was published in May 2026, co-authored by Takayuki Fukatsu, Kensuke Furukawa (Kensuu), and Kazuhiro Obara.
Meta-skills refer to the ability to decide "which task to give to AI" and "what to ask AI." It's not about how to use AI itself, but the power required on the human side using the AI.
The three authors each have different schools of thought; Mr. Fukatsu's is "Structural design that doesn't die." It's the idea of first designing a structure that survives in the long term while minimizing risk. From his experience with numerous products as a UI designer, he has arrived at the sequence of first creating a "system that won't break" and then running AI on top of it.
Learning how to write prompts is important, but that is a "tactic." Deciding where to place AI and what not to leave to AI is the "strategy." Mr. Fukatsu named the power to design that strategy "Meta-skills."
Prompt engineering might eventually become unnecessary as AI evolves. But the "ability to design systems that don't break" will remain no matter how smart AI becomes.
Method ⑥ AI Agents — Moving beyond chat to run autonomously toward a goal
Mr. Fukatsu is focusing on the evolution of AI from "one-on-one chat" to "agents that run autonomously toward a goal."
For example: "Research trends of top brands worldwide, extract brands with good compatibility, create proposal emails, and set up appointments." The AI executes this entire flow while making autonomous judgments.
Mr. Fukatsu particularly emphasizes compatibility with work where volume directly leads to results. Sending emails with individual text to a list of 1 million sales leads is impossible for a human, but possible for an agent.
On the other hand, he clearly states that "flows where humans and machines take turns passing the ball" are a poor match. If wait times or confirmation tasks are sandwiched in, autonomy cannot be utilized.
In this context, Mr. Fukatsu often talks about small and medium-sized enterprises in rural areas. A local factory can send proposals to brands all over the world through an AI agent. Language barriers and distance barriers disappear. "Generative AI and rural SMEs are an incredibly good match."
Method ⑦ Proactive AI — "It's finished before you even ask"
This is a concept Mr. Fukatsu discusses as a prospect for 2026.
Conventional AI utilization follows the order: "Human gets stuck → Ask AI → AI answers." This order will change.
Before the human even recognizes they are "stuck," the AI has completed the response in the background.
Method ④'s Calendar Secretary AI is the entry point. Getting information 15 minutes before a meeting is convenient, but it still assumes a "human registers the appointment." Beyond that lies a world where even registering the appointment becomes unnecessary.
Mr. Fukatsu himself says, "I want to experiment in 2026 with an AI that does my job for me." He admits it's not yet in its final form. But the direction is clear. AI is moving from "acting when asked" to "acting before being asked."
This isn't a far-off story; the entrance is already visible in Method ④. Whether you can skip the step of "asking because you're stuck" is the next move for those who have followed Mr. Fukatsu's seven methods.
Summary
Let's list Takayuki Fukatsu's seven methods once more:
- Method ① Fukatsu-style Prompt (Increase output accuracy with 4 elements)
- Method ② Make AI your "Master" (Use for human growth, not just tasks)
- Method ③ Management should be AI (As an unbiased outside director)
- Method ④ Calendar-linked Secretary AI (Automatic reports 15 mins before meetings)
- Method ⑤ Meta-skills (Train higher-level human skills before AI)
- Method ⑥ AI Agents (Run autonomously toward a goal)
- Method ⑦ Proactive AI (Finished before you even ask)
What all seven have in common is that they aren't just handing work to AI; they are using AI to change "human judgment."
It's not about memorizing a prompt template and being done. Beyond the template, there is a way to use it to grow yourself. Before eliminating tasks, there is a way to use it to change management decisions. Before running a single tool, there is the power to design "what to place where."
"Don't use AI for busywork." The weight of this sentence should sound different now than it did at the beginning.
If you try just one thing starting today, make it Method ②. Next time you talk to an AI, instead of saying "give me an answer," try writing "point out the weaknesses in my thinking." The quality of the words that come back will change completely.
Finally
What did you think?
One last thing.
Today's talk was only about "revisiting AI usage from a designer's perspective." However, there is a continuation that goes as far as "letting AI handle 80% of your work." I'll be blunt: I can't write all of that in a single article.
So, for a limited time only, I am giving away a set of "12 Major AI Utilization Benefits" for free.
From the basics of AI utilization to the system for delegating 80% of your work to AI. E-books, implementation guides, glossaries, and free individual consultations. I've packed all my current knowledge into these 12 items.
First, there are 6 e-books to build the "foundation" of AI utilization. If you read them in order from Vol. 1, you will naturally acquire essential AI skills.
- VOL.1 Prompt x Context Mastery (The two major elements that determine AI output quality in 5 chapters. Start here.)
- VOL.2 Task-performing AI Mastery (The entrance to "working AI" like Claude Code and Codex.)
- VOL.3 Knowledge Mastery (How to store and grow "information assets" that AI reads repeatedly.)
- VOL.4 Debugging Mastery (Practical techniques to turn "something's wrong" into "this is exactly what's wrong.")
- VOL.5 How to Delegate 80% of Work to AI (From inventorying tasks to API/MCP integration, rapidly expanding the scope of delegation.)
- VOL.6 Skills Mastery (Building a system where work ends with a single command. The final volume.)
VOL.1 is the direct continuation of today's talk. The four elements of the Fukatsu-style are exactly the design of prompts and context. If you've read this article, you can go straight into practice.
Next, 2 implementation guides you can start today. You can set up "working AI" on your computer without getting lost.
- Claude Code Implementation Guide (For Claude users. From installation to initial settings.)
- Codex Implementation Guide (For ChatGPT users. From installation to initial settings.)
Furthermore, 3 pages of AI terminology explanations with 60 words. All those words you're "too embarrassed to ask about" are here.
- AI Glossary ① | Basics (Generative AI, Prompts, Hallucination—20 terms about AI itself.)
- AI Glossary ② | Claude Code & Codex (Terminal, Skills, Path—20 terms for the screens you touch daily.)
- AI Glossary ③ | Integration & Tools (API, MCP, .env—20 terms you'll hit during automation.)
- And the 12th item:
- Free Individual Consultation (1-on-1 / X GROWTH). I will tell you one-on-one how to grow your AI presence on X as quickly as possible and earn ad revenue. It's a reproducible method even for absolute AI beginners, so if you "don't know where to start," please come. Slots are limited, so please hurry.
Total of 12 items. And they are all free. Even if you don't participate in the seminar or consultation, you can get all the e-books, guides, and glossaries.
Sounds too good to be true, right? But it's real.
To receive them, simply join the LINE Open Chat below.
Click here to start.
Let me say it one more time. What you need isn't technology. It's just a design that doesn't let the AI get lost.
Why not end the exhaustion of just aimlessly talking to AI today?
References
- "Management should be AI rather than employees" Mr. Fukatsu discusses 2026 Generative AI prospects and utilization strategies for local SMEs (CloudSign)
- A more efficient way than "using generative AI for work" Takayuki Fukatsu discusses business utilization of AI (Logmi Business)
- note, Takayuki Fukatsu appointed as CSO (Chief Strategy Officer) (PR Times, July 7, 2026)
- note will become the "content distribution infrastructure" of the AI era (note Inc., August 6, 2026)
- What is the Fukatsu-style Prompt? Explaining templates, 7 practical examples, and application points (SHIFT AI TIMES)
- Meta-skills: "AI x Self" Strategy in an Era Where the Value of Effort Changes Takayuki Fukatsu, Kensuke Furukawa, Kazuhiro Obara (NewsPicks Publishing, May 18, 2026)
- Mastering ChatGPT! The Fukatsu-style Prompt Reader Takayuki Fukatsu (Nikkei BP, August 2024)
