I have stopped "creating documents alone."
As a result, the hours that used to melt away into creating proposal materials have completely vanished. The time spent reading thick PDFs, reorganizing key points in my head, and staring at blank slides has been drastically shortened.
"Read everything first. Pick out the key points. Think of the structure. Write the text. Fix the formatting..." — doing all of this alone for a single document. Before you know it, half a day is gone from research to completion.
I have seen this exhaustion many times myself in the consulting field. The more talented people are, the more they tend to take on all the steps themselves and end up burnt out.
It's like a small shop where one person handles the kitchen, customer service, and the cash register. It's impressive to do it all, but if even one part gets stuck, the whole shop stops.
Document creation can be changed from "a job done entirely by one person" to "a job where roles are divided and rotated." Let the AI that is good at reading do the reading, and the AI that is good at creating do the creating. Just by doing that, those lost half-days will return.
Today, I will reveal the entire division of labor method, letting Gemini Notebook be the "Reader" and Claude be the "Creator," complete with prompts you can copy and paste. By the time you finish reading, you'll want to run your next document through this division of labor. Let's begin.
Chapter 1: Why Does Document Creation Consume So Much Time?
I want those who think "I'm inefficient because I'm incompetent" to read this. There are only three causes.
Cause 1: You are carrying all three steps of "Reading, Thinking, and Creating" alone.
Document creation is actually made up of three tasks with different natures: the process of accurately reading massive amounts of information, the process of assembling key points and thinking, and the process of creating it into a visual form. The way you use your brain is completely different, yet doing it all alone causes your brain to switch gears repeatedly, exhausting you every time.
Cause 2: Misinterpreting facts during the "Reading" phase.
Skimming a document in a hurry and starting to write based on vague memory. Later, someone points out, "That number is different from the source." If the accuracy of the reading is weak, the entire subsequent process collapses.
Cause 3: Starting from a blank page during the "Creating" phase.
You have the key points in your head, but when it's time to turn them into text or structure, your hands stop. A blank page is always scary. This startup cost is often the biggest time-sink.
What these three have in common is "continuing to hold onto tasks that could originally be separated." This is not a matter of ability. It is entirely a matter of method.
To be honest, I believe only those who have known the height of inefficiency are qualified to talk about efficiency. I once spent 5 hours on a single proposal document. Because I know that exhaustion, the effectiveness of separating tasks hits home.
Chapter 2: The Essence is to Stop "Holding it Alone" and Split into a "Reader" and a "Creator"
So, what should you do? The answer is "leave what they are good at to those who are good at it."
Please understand the difference in personality between these two AIs. This difference is the foundation of the division of labor.
The Reader: Gemini Notebook.
Gemini Notebook (NotebookLM) is an AI that answers based only on the documents you upload. Google's official help states that only sources are used for answers, and answers include citations (sources) so you can immediately check the relevant part of the original document. It is based on Gemini. It answers "I don't know" to things not written in the document. Therefore, it is hard to misinterpret facts. When it comes to "reading accurately," there is no better partner.
The Creator: Claude.
On the other hand, Claude is good at assembling text and turning it into document form. If you give it key points, it can think of the structure and create readable text. If you turn on "Analysis Tool" (code execution), it can even create Word, PowerPoint, or Excel files themselves.
According to official help, file creation is available even on the free plan. It defaults to the new Sonnet 3.5, and for more difficult tasks, you can switch to the top-tier Opus on a paid plan. It is the star of "thinking and creating."
In short, this is the division of labor: Let the Reader, Gemini Notebook, read the document accurately and extract key points with citations. Pass those key points to the Creator, Claude, to build them into a document. Let different professionals handle reading and creating. This is the overall picture today.
Chapter 3: Demonstration. How the Same Document Changes Between "Solo" and "Division of Labor"
Words alone are hard to convey, so I'll show you how it actually changes.
The task is: "Read a 30-page research PDF and create a one-page planning memo for internal sharing." A common job.
【When doing everything alone】
Open the PDF and read from the top. Highlight numbers that catch your eye. Summarize key points in your head. Open a blank memo and start writing. "Wait, where was that number?" and go back to the PDF. Write and fix, fix and write. Before you know it, 2 to 3 hours have passed.
—It gets done. But this is the result of repeatedly going back and forth between "reading" and "creating" in your head.
【When splitting into Reader and Creator】
First, put the PDF into Gemini Notebook and ask, "Extract the key points of this document with citations for evidence." What comes back is a list of key points with source numbers. Misinterpreting numbers is unlikely. Next, pass those key points to Claude and ask, "Assemble this into a one-page planning memo." The structure and text rise up all at once.
In my experience, the most effective part is that "brain switching disappears." The time to focus on reading and the time to focus on creating are cleanly separated. That stress of going back and forth is completely gone.
Same PDF, same you. The only difference is "whether you held it all alone or divided the roles."
Chapter 4: Practical Steps and Copy-Paste Prompts
From here, these are the actual steps. No programming required. No difficult integration settings; it works just by copying and pasting by hand. There are 3 steps.
STEP 1: Let the Reader (Gemini Notebook) read the document
Create a new notebook in Gemini Notebook and add the documents you want it to read. The trick is to summarize one theme per notebook. Gemini Notebook cannot span multiple notebooks simultaneously, so mixing them reduces accuracy.
STEP 2: Extract key points with citations
Paste the prompt below into the Gemini Notebook chat.
1Please extract the key points of the uploaded documents in the following format:21. Conclusion of this document (within 3 lines)32. Numbers effective for decision-making (each with a source citation)43. Parts that could be counterarguments or points of caution
For things not written in the document, do not supplement with guesses; clearly state "Not mentioned."
Key points will return with sources. Always check the source before using the numbers. This finishes the "reading."
STEP 3: Let the Creator (Claude) assemble it into a document
Copy the key points you just extracted and give them to Claude.
1Below are key points extracted from a research document.2Based on this, please create a planning memo (equivalent to one A4 page) for internal sharing.34【Key Points】5(Paste the output from STEP 2 here)67Make the structure: "Conclusion → Evidence → Next Steps."8Do not add facts not written in the key points, and clearly state "To be confirmed" for missing information.
With this, a readable page is created. If you add "Make it a Word file," it will even output the file.
One important thing here: this division of labor is a path I have walked myself. As a result of building up prompts exceeding 20,000 characters to create my own dedicated AI, the time taken for one article went from 5 hours to 1 hour. It's the same AI. The only thing I changed was how I handed things over. Deciding who to entrust with what. That alone changes how time melts away.
Chapter 5: 3 Applications to Get 10x Out of This Division of Labor
Once you have the basics, mastering these will take you to the next level.
Application 1: Return the created output to the Reader to grow it
Add the memo created by Claude back into the Gemini Notebook sources. From then on, that output also becomes a "target for reading." The Reader's knowledge thickens every time you use it.
Application 2: Let it read multiple documents at once
If you put several related documents into one notebook, Gemini Notebook can pick up key points across them. Methods like "Compare the documents of Company A and Company B and show the differences with citations" are effective.
Application 3: Integration to eliminate back-and-forth (For advanced users)
For those using Claude Desktop or Claude Code, you can use a mechanism called MCP to connect directly with Gemini Notebook, eliminating even the back-and-forth of copying and pasting.
However, to be honest, this is for advanced users. This integration is not officially guaranteed by Google or Anthropic, and it may stop working if the screen layout changes. Using it also requires a paid Claude plan. In the beginning, don't force a connection; manual copy-pasting is sufficient.
Finally, a deeper look at why this division of labor works. As AI becomes smarter, the next bottleneck won't be "the power to create." It will be how to get the receiver to accept what was created. People don't move easily when handed a finished product out of the blue. This division of labor, where the Reader leaves evidence with citations, works exactly there. You can hand it over with the evidence, not just the finished product. That's why it passes easily.
Summary: Document Creation from "A Job Held Alone" to "A Job Handed Over"
This time, I explained the division of labor of entrusting "reading" to Gemini Notebook and "creating" to Claude, from the structural causes to demonstration, steps, and applications. Finally, the core of this article once more.
Document creation gets bogged down not because you are incompetent. It's because you are carrying jobs with different natures—reading, thinking, and creating—all by yourself. Let the AI good at reading read, and the AI good at creating create. The moment you separate the processes, that half-day quietly disappears.
That said, you don't need to do everything at once. There is only one thing to do.
Just put one document you have on hand into Gemini Notebook and try extracting the key points with the prompt from STEP 2 in Chapter 4.
Once the key points with citations come back, just hand them to Claude. Once you taste this feeling of "reading and creating being separated," you will never be able to go back to the days of holding everything alone.
Both Gemini Notebook and Claude are evolving very fast. Rather than trying to keep up with the latest alone, it is overwhelmingly faster to learn from those who are practicing it.
Thank you for watching until the end.
【📣 Notice 📣】
I am opening an Open Chat to completely master AI × note and X.
In the Open Chat, I deliver the latest and useful information on AI × note and AI/X without reservation.
🎁 Open Chat Free Benefits 🎁
① 200 Selected Prompts
② 20 Selected Gems
③ 7 Selected Claude Skills
I am giving these as presents 🎁
🌈 Content Shared in Open Chat 🌈
① How I achieved 1 million yen in sales with my first-ever note
② SNS marketing methods
③ List marketing methods
④ Digital data marketing methods
⑤ How to grow on X at the fastest speed
And more, I share information I can talk about from my experience as an active AI × SNS marketer who previously did digital data marketing and big data marketing.
AI beginners and those who just want to watch are also OK ✨
Please feel free to take a look ✨
↓ Please join from here.
https://line.me/ti/g2/LmLu1N1cE6UBkoURbaYf_bV8l66cCyotSJU2og
