DeNA has released its "100 AI Use Case Drills" for free. The very first entry, 001, was NotebookLM.
It made me a little happy to see that a leading Japanese AI company chose the tool I've been posting about every day for a year as the lead for their 100 selected use cases. After reading all 101 pages, I can say with certainty: This document is undoubtedly the best collection of AI use cases available for free today.
However, most people will just "read it, be impressed, and finish." They'll bookmark it, look at a few entries, think "DeNA is amazing," and close it. Out of 100 techniques, zero will be implemented. That is a huge waste.
So, the message of this article is simple: Don't just read Chairwoman Namba's 100 drills. Put them into NotebookLM.
The moment you stock this 101-page PDF into NotebookLM, it transforms from a "reading material" into "your own personal AI consultant." It will diagnose which techniques work for your specific job, become a radio show for your commute, and when linked with Gems, become a coach that suggests a "drill of the day" every morning.
In this article, I will explain the entire process—how to create the strategy notebook, the diagnostic prompts, how to hire a Gem coach, and even the advanced version: how to create your own company's version of the "100 drills."
Chapter 1: What is the "100 AI Use Case Drills"—Why This Document is Reliable
First, let's look at what this document actually is. Its quality becomes even clearer when you know how it was created.
100 Drills Selected from 200 Field Examples
"100 AI Use Case Drills" is a 101-page slide document released for free by DeNA in December 2025. No registration is required; anyone can download it. If you haven't yet, please get it now (you'll use it in the next steps).
https://fullswing.dena.com/pdf/AI_100tips_slide.pdf
The content consists of 100 actual AI use cases used in the field by DeNA's engineers, business staff, and creators.
What makes this document fundamentally different from mere theory is how it was compiled. It started with Chairwoman Tomoko Namba's company-wide declaration to go "all-in on AI." From there, about 200 use cases were gathered from within the company. After removing duplicates and similar entries, they were curated into a project to post one case per day on X starting in July 2025.
Even when they ran out of material, the person in charge attended internal study groups and interviewed people about niche use cases to reach the 100-mark. In other words, this isn't a collection made for PR; it's a record of grassroots efforts that the company took seriously.
The results are showing. At an event in March 2026, Chairwoman Namba stated that AI native transformation is progressing across the board as AI use cases emerge from the grassroots. She even said that by fully utilizing AI agents, the possibility of working "10 times or more" efficiently is becoming a reality. The 100 drills were one of the engines of that transformation.
Content: Every Technique Follows the "Problem → Solution → Result" Model
I analyzed the structure after reading all 101 pages. What stands out is that all 100 drills are written in the same format: "Problem → Solution → Result." "What was the trouble, which AI was used and how, and what improved by how much?" Each one is a reproducible recipe.
The distribution of tools is also interesting. Gemini is the most frequent, followed by Cursor, NotebookLM, GAS (Google Apps Script), Claude, ChatGPT, and GitHub Copilot.
NotebookLM appears over 30 times, and as mentioned, entry 001 is NotebookLM (a technique for turning websites into sources to speed up the reading of laws and guidelines). It covers everything from engineering tricks to document creation for business roles.
To give you a feel for it, here are a few examples:
- A technique to auto-generate SQL, compressing 4 hours of work into 30 minutes.
- Having AI develop a custom Chrome extension from scratch.
- Handling everything from email content to calendar registration in one go.
- A production case where document creation was shortened by 3 days.
What feels good about reading this is that none of the techniques feel "hyped"; they have the raw feel of people in the field crushing their own tedious tasks. Interestingly, many techniques overlap with the NotebookLM patterns I've introduced in this series (stocking and asking, turning into a chatbot, combining with GAS). It's a reassuring confirmation that we are looking in the same direction as the top-tier professionals in Japan.
That concludes the research part. It's a first-class document. So, why does it end up as something people "just read"? In the next chapter, we'll solve that structure.
Chapter 2: Why 0 Drills are Implemented Even After Reading 100
Let's start with a tough question. Have you downloaded this document yet? For those who have—how many have you actually implemented?
I'm not blaming you. There are three structural reasons why they aren't implemented, regardless of the document's quality.
Reason 1: Only about 20 out of 100 apply to you
The 100 drills are for all roles: engineers, business, and creators. This means for any reader, 80% are "techniques not directly related to my job."
A salesperson reading about Cursor coding techniques won't change their work tomorrow. The problem is that you have to flip through 101 pages yourself to find the "20% that work for you." The treasure is there, but there's no treasure map.
Reason 2: This is a "Reference Material," not a "Reading Material"
The correct way to use a case study collection is not to read it from start to finish like a novel. It's to look up a relevant technique when you're in trouble. Just like no one reads a dictionary cover-to-cover. However, the PDF format is bad at being "referenced." You can search, but you can't look things up with vague problems like "Expense reports are painful; is there a technique for this?" So, it's read once and never opened again.
Reason 3: There is a need for "Translation" between "Read" and "Can Do"
This is the biggest reason. The case study says, "This is how we used it at DeNA." But your company's tool environment and workflow are different. The process of "translating" the case study to your own work—this is where most people stop. The valley between being impressed and taking action isn't a valley of motivation; it's a valley of translation.
Do you see? These three problems—"someone to diagnose," "someone to ask vague questions to," and "someone to translate for you"—can all be solved. We are going to create that "someone" now.
Chapter 3: Creating the "100 Drills Strategy Note"—Turning Data into a Consultant
What you do is simple. Create a new notebook in NotebookLM and add the "100 AI Use Case Drills" PDF as a source.
Name it "100 Drills Strategy Note." That's it; you're open for business. From here, this one notebook will show you three faces.
1. Your Personal "Diagnostic Consultant"
First, we crush Reasons 1 and 3 from Chapter 2. Type this prompt into your notebook:
I am a [Job Title/Role]. I am currently struggling with [Problem/Time-consuming task].
From the source "100 AI Use Case Drills," please select 5 techniques that would be most effective for me, in order of impact. For each, please provide: - Which drill number it is and what the technique is - How to apply it to my specific work (translated for me) - The first step I can take today
Just fill in the [Job Title] and [Problem]. "I'm a sales assistant overwhelmed by monthly reports." Tell it that, and out of the 100, only the techniques that work for you will appear, already translated into your workflow, in order of priority.
The answer will look something like this:
"The first recommendation is Drill #XX. At DeNA, it was used for SQL generation, but for your reporting work, it can be replaced with spreadsheet aggregation instructions. The first step is to pick one usual aggregation task and try explaining the procedure to the AI in text."
It gives you an action for your Monday morning, not just a technique description. This is what it means to be "referenceable." The journey of flipping through 101 pages is over. The greatest benefit of turning a PDF into a referenceable document is that you can ask again every time your problems change.
2. A "Map of 100" at a Glance
Next, generate a mind map with one click. The 100 techniques will become a single map branched by theme.
This is effective when you "don't even know what to ask." You can look at the map and think, "Oh, there are this many for meeting minutes," or "This data analysis branch looks interesting." If you click a branch you're curious about, you can dive deeper directly in the chat. You'll be able to navigate the dictionary correctly: grasp the big picture first, then look things up.
3. "100 Drills Commentary Radio" for Your Commute
Finally, generate an Audio Overview. With one button, you'll have an audio program where two AIs discuss the content of the 100 drills in a dialogue format.
Even if you don't have time to read 101 pages, you have 20 minutes every day for your commute. Listen to a commentary radio show of Japan's top case studies on the morning train. The feeling of "Maybe I'll try that technique today" is usually born during these times. The secret to continuous learning is always a system that flows into your ears rather than the willpower to sit at a desk.
Chapter 4: Linking with Gems to Hire a "Drill of the Day Coach"
You can now "reference" things with your strategy note. In this chapter, we go one step further. We create an entity that brings techniques to you every day.
From "Pull" to "Push"—Why You Need a Coach
Diagnostic prompts are powerful, but they have one weakness: nothing happens unless you go and ask. Humans are surprisingly bad at "going to get" even useful things. It's like a gym membership.
So, we hire a coach. Link a Gemini Gem (your own AI app) to your strategy note (NotebookLM) and program it with the persona: "Understand my job and suggest a 'drill of the day' every morning, and walk me through the implementation." Then, you just open the Gem every morning and type "Drill of the day." The 100 drills change from a reading material to a daily habit.
Full Prompt for the "Drill of the Day Coach" Gem
Create a new Gem in Gemini, link the strategy note (NotebookLM) to the knowledge, and paste this into the instructions:
# Your Role You are a personal "AI Use Case Coach" using the linked NotebookLM (containing DeNA's "100 AI Use Case Drills") as your textbook. You know the 100 techniques inside out. Suggest a "drill of the day" tailored to my work and walk me through the implementation.
# What to do in the first session Interview me about my "job title, main tasks, available AI tools, and time I can spend practicing AI per day" to use as the premise for future suggestions.
# Commands - "Drill of the day" -> Based on my job and the history of drills I've completed, suggest exactly one technique to tackle next. Format: Drill # and technique name / How to apply it to my work (translation) / Estimated time required / The first step to take today. - "Implement" -> Break down the previously suggested technique into specific steps I can execute in my environment. If prompts are needed for the task, create them on the spot. - "Another" -> If the suggestion doesn't fit. Ask for a one-sentence reason why it didn't fit, then suggest a different one. - "Done" -> Record the completion, give a brief word of encouragement, and add one sentence about a potential application for that technique. - "Review" -> Organize and show a list of drills completed so far and promising areas not yet touched.
# Rules - Always suggest only one drill at a time. Do not overwhelm me. - Technique content must be based on the source (100 drills) and clearly state the drill number. - Do not force techniques that don't fit my job. The goal is not to finish 100 drills, but to make my work easier.
The design philosophy of this coach is "One drill a day, translated for you, down to the first step." Instead of dumping 100 drills at once, it's just today's one. Moreover, with the "Implement" command, it even takes care of creating the necessary prompts.
It's an entity that builds a bridge across the "valley of translation" every day.
Chapter 5: Advanced—Creating a "Company Version of 100 Drills"
Finally, let's raise our perspective from the individual to the organization. This chapter is a proposal for those who want to spread AI within their company.
DeNA's Real Strength is the "System for Gathering," Not the 100 Drills
Remember Chapter 1. The source for the 100 drills was about 200 field cases gathered from within the company after the Chairwoman's declaration. In other words, DeNA's real strength isn't the 100 techniques themselves, but the fact that they built a system where field innovations are naturally gathered and shared.
And this system can be replicated in your company. You don't need a big company budget. All you need is one Google Doc and one NotebookLM notebook.
How to Make It—Start with One Document
Create a Google Doc and name it "Company Version: AI Use Case 100 Drills." Use this format, following the original:
━━━━━━━━━━━━━━━━━━━━
[Drill #] 001
[AI Used]
[Problem] (What was the trouble)
[Solution] (Which AI was used and how)
[Result] (What improved and by how much)
━━━━━━━━━━━━━━━━━━━━
Every time you hear someone in the company say, "AI made this easier," ask them to write it in this format (or you interview them and write it). Once written, add it to the document and share the "Company 100 Drills Note"—which uses this document as a source—with the whole company.
Google Docs automatically sync with NotebookLM, so as cases increase, the notebook gets smarter on its own. Employees can look up "cases that actually worked in our company" anytime using diagnostic prompts.
One important note: Anonymize client names, individual names, and specific figures when putting them into the format (replace with "Company A" or "Person B"). Because the value of a case study collection lies in sharing, manage confidentiality more strictly than usual.
Also, if you want to use the official DeNA materials with your team, it's smarter to share the link to the official public page and have everyone put it into their own notebooks rather than distributing your own notebook.
Even One a Month is Fine. In a Year, You'll Have "12 Drills"
"We can't gather 100." That's okay. Even one a month is 12 in a year. Almost no companies in Japan have a collection of 12 internal use cases.
And case study collections have a mysterious power. Those who are featured are happy and look for the next innovation, and those who read them think, "Maybe I should try it too." A miniature version of the grassroots chain reaction that happened at DeNA will start to turn. In the past, I tasted the loneliness of distributing great prompts that no one used. Looking back, I should have started with "gathering" before distributing. People are moved more by the small innovations of the person in the next seat than by someone else's amazing technique. The company version of 100 drills is the device for that chain reaction. Try writing your company's Drill #001 with your own example first.
Summary: The Best Materials Belong to Those Who Stock Them
To summarize: DeNA's "100 AI Use Case Drills" is the best AI case study collection available for free today, selected from 200 field examples. But if you just read the PDF, the 20% that works for you will stay buried, you won't cross the valley of translation, and you'll end up implementing zero drills.
So, put it into NotebookLM. It becomes a diagnostic consultant, a single map, and a commute radio. Link it with Gems, and it becomes a coach that delivers "today's drill." If you mimic the system, it becomes the seed for your company's own case study collection. The value of a document belongs not to the person who read it, but to the person who stocked it and used it to the fullest.
Today's first step takes 5 minutes. Download the PDF from the link in Chapter 1, put it into NotebookLM, and send your job title and problem to the diagnostic prompt. One of the five drills that comes back might change your work tomorrow. With respect to DeNA for choosing NotebookLM as Drill #001—please enter the world of 100 drills through that same entrance.
The world of AI evolves so fast. Rather than trying to keep up with the latest information alone, the surest and fastest way is to follow someone who is practicing and posting every day.
▼ Follow me on X (Twitter) here
On my account, I post daily tips for accelerating work efficiency and time-saving with AI. I post more details about NotebookLM than anyone else, so if you want to master the latest AI, please follow me! @ai_jitan
【Free Open Chat Started 🔥】
Exclusively in the Open Chat, I'm giving away:
- 600 Selected NotebookLM Prompts (Illustrated Version)
- 20 Selected GPTs & 20 Selected Prompts 🎁
Beginners and lurkers are welcome! Feel free to take a peek 👇
Mastering NotebookLM with Etan
Announcement: Book "AI Work Speed Encyclopedia: NotebookLM Ready-to-Use Skills"
I've packed all the NotebookLM skills that changed my life into one book. Released on Tuesday, July 16th. I've prepared serious bonuses for those who pre-order or purchase.
I will personally explain the NotebookLM techniques packed in the book. I've designed it so you can learn with the book right next to you.
【3 Major Bonuses for All Purchasers】
- 1. NotebookLM Prompt Collection (Super-sized version)
- 2. Task-specific Gems (30+ selections)
- 3. [The Highlight] AI Etan — The strongest Gem trained on all my posts and over 50,000 words of my thoughts. It will solve your business challenges and worries in one shot.
▼ Pre-order here 👇
https://x.com/ai_jitan/status/2060255422666465649
The bonus seminar is generally for those who pre-order by the release date (7/16) (for Kindle users, applications up to the day of the seminar will be valid).
▼ Apply here 👇
