Chapter 0
What happens after a company adopts AI
Why an AI initiative left the company working the same way
A proposal requested on Friday arrived first thing Monday, finished by a junior employee and AI while the person in charge was away. The CEO announces that the company will adopt AI; the employees exchange uncertain looks.
Tamura is 55 and has never used AI. He searches for “What is AI?”, glances at a colleague's screen full of code, and quietly looks away.
Sato, a salesperson, occasionally asks AI to improve an email. The wording is better, but preparing the quotation will still take three hours; the work itself has not changed.
Colleagues praise Nakanishi, the digital transformation lead, for building an automation. In private, he admits that the company's way of working is still the same.
Following the CEO's announcement, Sato tries preparing quotation materials with AI. He keeps correcting the instructions, including which price table to use.
The quotation is finished, but repeated instructions and corrections take almost as long as the original work. AI has been working, and Sato has been working the entire time too.
Aoki in planning uses AI every day, yet starts each request by explaining the background from scratch. Even a long explanation does not reliably produce the result she intended.
Aoki posts a price change in chat and sees read receipts. Days later, Sato is surprised to learn about the change: nobody has gone back to check the older message.
Aoki reminds Sato to use the usual format for delivery dates. She gives the same explanation the next week, and again the week after.
The four colleagues sit exhausted in the break area. Each has a different problem, while the CEO calls from the corridor to ask how the AI initiative is going.
Three colleagues visit Yamaguchi Manufacturing, the company that finished the proposal without a handoff. Staff enter a one-line request in the relevant work record, with no separate explanation or transfer of documents.
The request arrives alongside the full discussion history, documents and current state, so AI can read them and begin. Nakanishi is especially surprised: AI is making decisions, beyond executing a fixed procedure.
The visitors see no one preparing status reports and no status meetings on the calendar. Everyone can check progress whenever needed by opening the Goals.
An employee explains that AI assessed and handled three items overnight; only one needs human attention that morning. Nakanishi realizes that the team delegates decisions as well as routine work.
Each work record contains the context needed for anyone to continue. The tools are the same; the team changed how it organizes work after taking a course on the structure of work.
Back at their own company, everyone takes the same course, including the CEO and department managers. They begin by changing how work and its context are organized.
Aoki no longer explains the same background every time, and Tamura can get work moving by asking AI to continue preparing an invitation. The CEO checks progress on screen, and the misunderstandings in meetings disappear.
Aoki can hand work to Sato with one request to take over the Goal. The training they took is the course you are reading.
The underlying change has happened in practice. At Addness, keeping Goals and their current state together reduced a handoff that took a full day to 30 seconds. The story is published in an example on note.
What you will be able to change
You do not need a clear picture of everything AI can do. Having asked ChatGPT a question is enough to begin.
First, the amount of manual work decreases.
| Work | Before | After |
|---|---|---|
| Customer inquiries | Read and answer each message as it arrives | Sorting and standard replies run automatically; only messages that need judgment remain for you |
| Articles and posts | Choose a topic, research it and write it | AI selects a topic and prepares a draft |
| Proposals and documents | Start from scratch each time | Ask “Continue the Acme project through to a proposal”; AI reads the background and prepares it |
| Collecting and sharing numbers | Collect data on Friday, copy it into a report and write commentary | The report arrives at 5 p.m. on Friday with reasons for the changes |
| Sorting and recording documents | Name each document and add it to a register | Filing and data entry happen automatically; uncertain cases remain for review |
The course goes beyond making individual tasks faster.
Speed alone also makes unnecessary work faster. Producing 1,000 documents overnight will not change the company's results if nobody uses them. The course spends most of its time on the changes that follow.
| Area | Before | After |
|---|---|---|
| The current state of work | Someone must ask, “How is that project going?” | Ask AI and get an answer immediately. “What is the status of the Acme project?” returns the current state and the next action |
| Reports and meetings | Prepare documents to share progress, then meet to discuss them | Status documents and reporting meetings become unnecessary. Meetings remain for decisions |
| Handoffs | Meet, explain the history and pass on the documents | Make one request to AI. Work continues without a separate explanation or document collection. A day-long handoff took 30 seconds |
| Stalled work | Work stays stalled until someone notices | On Monday morning, stalled work is listed with the reasons it stopped |
| How people use AI | Everyone explains the background from scratch to their own AI | Everyone's AI knows the company's current situation. Repeated explanations are unnecessary |
The first table describes steps toward the second. The purpose is to change how work proceeds across the company.
Why go that far? Consider what becomes possible when AI has as much context as a person doing the work.
Today's AI is capable enough that, given the right information, it can provide more value than hiring one highly capable person. It can also do a great deal of work at any time, at a much lower cost than such a hire.
What it lacks is knowledge of what is happening inside your company. The course explains how to provide that knowledge, from beginning to end.
The company in the story struggled because of how it organized work. That is why the course starts with the structure of work.
How to take the course
After the introduction, there are six core chapters, a setup guide and one operations lesson. Understand the structure of work, make it possible to delegate to AI, prepare the working environment, run work automatically, and extend the approach to the team. Follow the sequence.
- Chapter 1: Five stages of getting results with AI — Begin with the five stages.
Tool-specific guidance, including Claude Code setup, AI behavior and automation examples, appears in supplementary lessons. Open them alongside the core course whenever needed.
- Supplementary lesson 1: A complete guide to Claude Code
- Supplementary lesson 2: A complete guide to Codex
- Supplementary lesson 3: Examples of automation and autonomous work
- Supplementary lesson 4: Understand how AI behaves
- Supplementary lesson 5: A glossary for working with AI
Unfamiliar terms are explained in Supplementary lesson 5: A glossary for working with AI. The quickest way to get an explanation is often to ask AI. A question such as “What does context mean?” is enough.