From here, I’ll introduce a concrete development method.
Exploratory is, as the name suggests, an exploratory method: you set an exploration period, and during that time, each person investigates relentlessly on their own.
There is no plan, no deadline, and no meetings or conversations. If the exploration period is two weeks, for example, each individual spends that time freely exploring on their own. However, this alone has limits, so each person must always publish their exploration process and results. Japanese engineers may be familiar with “funpou” (if not, imagine someone posting many times a day on X)—that’s roughly the level of broadcasting expected. You must do it. By doing so, you can get hints and inspiration from other members (and the information they share).
Exploratory follows these three steps:
Let’s go through them in order.
In the Preparation Step, you decide the exploration period, members, theme, and workspace.
For the exploration period, you want at least one week. Ideally, two weeks or more. Keep the team to 4 people or fewer, and have them commit as fully as possible. At minimum, each person needs to spend more than half a day on it. This also means securing that level of resources (including labor costs). Saras Sarasvathy’s Effectuation has the “Affordable Loss” principle—decide in advance the maximum you can afford to lose, even in the worst case. It’s the same idea. This starting line may sound strict, but treat it as a prerequisite. Exploration requires deep focus, and because there is no right answer and you must keep making decisions yourself, it’s easy to become exhausted—so you need at least this much in place. If you can’t get to this starting line, start by making it possible to stand on it.
As for the theme, specify in a word or two what kind of soft skill you want (or should) create. It’s fine if it’s rough. As you explore, you’ll inevitably find multiple candidates and later be forced to choose among them. If anything, at this stage you should not define it in more detail than a word or two.
Finally, the workspace is where all members share their exploration process and results. Options include a wiki like Cosense, a note tool like Notion, or a repository like GitHub. Whatever you choose, make sure it can handle 10,000 characters of text without friction. In practice, this means three options: share in plain text via GitHub, split it up in a real-time collaborative app like Cosense or Notion (e.g., 2,000 characters × 5 pages), or use both. Put differently, email, chat, and cloud storage are out of the question. With those tools, the cost of sharing is too high, and high-frequency sharing won’t hold.
I wrote “high-frequency sharing,” but that’s misleading. It’s more than that. Ideally, it would be always-on sharing—sharing essentially everything you touch and everything you write during exploration. But you can’t livestream 24 hours a day, and that would be too much information to be useful. That’s why the practice of frequently coming to the workspace and sharing is, in effect, indispensable. You can write dozens of times a day like an X addict, or you can post about once every half day.
In any case, you must share—always. And not as an unfiltered dump: make it readable for other members (or curate it). This is called the sharing obligation.
The sharing obligation is the core of Exploratory. In exchange for having no planning, management, or meetings at all, members rely on what’s shared. Solo exploration has limits, but if you can freely browse how others are exploring, you can gain hints and inspiration. That makes it possible to keep going even in lonely exploration.
In the Exploration Step, you explore until the exploration period ends.
What should you do as exploration?
There is no answer. Other than using “the theme decided in the Preparation Step” as a starting point, you have complete freedom. You can do thorough solo brainstorming, dig into papers nonstop, or focus on interviews and hearings. Strategically, you can explore broadly and shallowly first and then head toward what truly matters once you find it, or you can pick a single direction from the start and persist in it. There is no single right approach. During the exploration period, mobilize everything you have—expand. Dig deep. And share.
However, you must always create an exploration summary. This is the deliverable of Exploratory, and it consists of the following three things:
As stated earlier, soft-skill development is about creating concepts, so some concepts should emerge. Depending on the period, dozens or even hundreds is not unusual. Also, just as software is structured into blocks of lines, functions, and modules, soft skills should be structured from smaller concepts to larger ones. Let me restate that: structure them.
Typically, you define a soft-skill tree and place the concepts into it as components, but it doesn’t have to be strict, and it doesn’t have to be a tree—just structure it. Finish it so it can be understood from a whole-system perspective as well. Skipping this is like saying you built only a library in software development. What you should be building is a product. Likewise, in soft-skill development, you want to create a soft skill—you don’t just line up individual concepts and stop. You must decide for yourself how to combine those conceptual components and what they achieve.
Finally, for context: organize the background that led you to include those concepts and that structure in the summary. References are easy to understand, but what we really want is design intent. Write out why you did it, what you did, and what twists and turns you went through. Since a generative AI can interpret it, focus on packing in information without omissions rather than making it clean and compact. It’s like an appendix; it can be massive—tens of thousands or even hundreds of thousands of characters. However, at minimum, organize it well enough that other members could make sense of it even without using generative AI.
The Reflection Step is conducted after the exploration period ends. All members report the situation and discuss what to do next.
As a prerequisite, create the exploration summary in the previous Exploration Step. Also, as much as possible, it helps to think through your own view on the next action (described later). In other words, you don’t start summarizing during the Reflection Step; you summarize before the Reflection Step. In the Reflection Step, you want everyone to be able to read each member’s summary.
As for the format, I recommend an environment that includes staying overnight—like a camp or retreat—where you can focus on the Reflection Step all day. Rather than cramming in a curriculum like a training program, a retreat where each person can spend about half the time freely is better. After all, you’ll likely be exhausted after exploration, and your perspective may have narrowed. Letting off steam helps you refresh both mentally and physically, which tends to improve the quality of discussion and decision-making.
As for reporting as well, I don’t recommend formal facilitation like having each person present and then doing Q&A (though this approach may fit some teams). Instead, a “silent co-working” style is preferable: everyone is gathered, but no one talks; each person quietly reads the exploration summaries, while talking only as needed. Exploratory relies, from start to finish, on deep individual focus. Reflection is no exception. Deep focus is difficult in face-to-face verbal conversation, so it should be avoided as much as possible. As an exception, if all members are strong at verbal communication and psychological safety is ensured, then verbal discussion is fine.
The detailed procedure for the Reflection Step doesn’t matter. You only need to meet the goal. The sole goal of this step is to decide the next action. Concretely, there are these three patterns:
If you abort, it’s over for now. Formal organizational activity is stopped.
If you retry, you loop back to the Preparation Step again. You should revise the theme, and you should either not carry over the deliverables so far, or only reference them as needed. Exploratory hinges on deep, focused individual thinking, and past deliverables interfere with that. If you want to retry in the first place, it means the exploration isn’t going well, and you need to reset with a flexible mind.
The only exception is when you simply didn’t have enough exploration time. If you didn’t have enough exploration time, it’s OK to just extend the period under the same conditions. Resources aren’t unlimited, so typically you’ll add short exploration periods incrementally. In other words, rather than setting a six-month period from the start, it’s usually more realistic to do one month, then another month, and so on.
Finally, if you commit, draft a project proposal. Literally, “commit” means to promise/declare or to go all-in, but here it means what’s beyond development—Action in the ABCD Walk. Exploratory covers development of the soft skill, but not subsequent submission and operation. You likely want to actually influence someone; but to do that, you need an appropriate stage—formal organizational activity—which will probably take the form of a project. Even if you don’t have the role or authority yourselves, you should still be able to create a proposal for those who do, and plan an approach strategy.
Here are a few tools for running Exploratory.
If you want to start Exploratory, you want the number of “No” answers to be zero. If you have any “No,” I recommend doing the research, consideration, and practice needed until they become “Yes.”
# Exploration period
- Is the exploration period 2 weeks or longer?
- y | n
# Members
- Are there 4 members or fewer, including you?
- y | n
- During the exploration period, can the members spend more than half a day exploring?
- y | n
- Likewise, have you secured the time, financial, and cognitive resources needed to spend that much?
- y | n
# Theme
- Is the theme decided?
- y | n
- Does the theme fit within 2 lines and within 100 characters (or within 200 characters in English)?
- y | n
# Workspace
- What tool do you plan to use as the workspace?
- (Write)
- If you use multiple tools, does it fit within two: one main tool you read/write in frequently every day, plus one supplementary sub tool?
- y | n
- If you use multiple tools, can exploration still work with just the main tool you read/write in frequently every day?
- y | not sure until we try | n
- Do all members belong to the workspace and have permissions that allow them to read and write without restriction?
- y | n
- Does the workspace have no notifications, or can notifications be disabled via settings or operational workarounds?
- y | n
- The workspace does not include email, chat, or cloud storage
- y | n
- Is there a clear path that ensures all members will definitely come and check the workspace?
- y | n
- Is that path something you can keep checking every day for over a month without getting tired of it?
- y | not sure until we try | n
A trigger list (a set of perspectives) is something that prompts ideas and action just by looking at it. Exploration is a completely free and autonomous activity, and until you get used to it, you may not know what to do—so it’s useful to store hints for deciding what to do.
# Exploration time
- What time of day will you explore?
- Can you secure at least 2 uninterrupted hours without interruptions?
- If you can't, what will you do instead?
# Exploration balance
- How do you want to allocate input, thinking & trying, and output?
- Why do you want that allocation?
# Between input and output
- Do you think what's between input and output is "thinking and trying"?
- Then what about "putting it into words"? Does that feel right?
- If it's neither "thinking and trying" nor "putting it into words," what do you think it is?
# How to do it
- Input sources
- Other members, internal, internal
- Books, papers, online documents, articles
- AI chat, Deep Research
- Desk research, hearings, interviews, etc.
- How to think and try
- Creative thinking (diverge → converge → distill)
- Soundboarding: with yourself (self-questioning), with others, with AI, etc.
- Prototyping: illustrations, software, other works
- Classify existing information using existing frameworks or your own, etc.
- How to output
- Write frequently about what you did, like funpou (high-frequency posting 10+ times/day)
- Create 1 to 3 checkpoints per day and write a summary there
- Scribble as unorganized notes, then have generative AI summarize at checkpoints and publish it
- Roam the workspace and post as reactions to other members' writing, etc.
# Monitoring exploration
- Can't get yourself to start exploring? If so, what's the cause? How should you address it?
- Do interruptions occur during exploration? If so, why? What should you do to prevent them?
- Are you stuck on something? If so, what's the cause? How can you remove the cause?
# Analyzing exploration time
- Do you see any patterns in when you explore? If so, why do those patterns occur?
- Is that time slot appropriate for you? Would it be better to change it?
- Are you for or against fixing your exploration time slot? Why?
# Exploration summary
- How will you build the exploration summary?
- What concepts will you create?
- How will you structure those concepts?
- How will you organize the context that led to those concepts and that structure?
- What next action does that exploration summary assume?
- Abort (stop)? Why?
- Retry (re-explore)? Why?
- Commit (exploration is done, transition to formal activity)? And what kind of activity will you do?