We don't stop at "training done." We stay beside you until it's usable.
A single day of training won't leave people merely thinking they've got it. We coach beside your employees until they can use AI in their own work — and keep using it.
Even if the groundwork is set from an assessment, "keeping it in use" is a different matter.
You ran the training. But by next week, is everyone back to the old way?
Many companies get stuck at the same place. We were the same at first, too. At month-end, HR would "disappear" for two full days, and a new hire's questions chopped up a senior's focus every day — that was the kind of company we were.
What's missing isn't motivation — it's a way to keep it in use.
People who work with AI every day teach it, fitted to your actual work.
Not general theory learned elsewhere, but the methods we've proven in our own real work, translated to your company's jobs.
The people who teach it use it across our whole company first.
01 AccountingLogging receipts
Typing in → reading the numbers
beforeWe used to open each invoice and receipt PDF one by one, decide the journal entry, and type it into the accounting software by hand. 120 minutes a month for one person.
nowNow AI gathers receipts from 14 services and employees' receipts and drafts the journal entries. A person just checks and hits ✓. It's down to 35 minutes a month.
02 Sales (junior)Prospect lists and inquiries
A user → someone who removes their own busywork
beforeA new hire hand-built a list of 20 companies × 5 sectors = 100 prospects, then retyped the same boilerplate — changing only the addressee — into each company's form. That alone burned a whole day.
nowWe handed AI to a 22-year-old in sales, and they turned list-building through sending into a system themselves. 100 companies now take about 105 minutes.
03 HRMonth-end tallies and pay statements
Two days gone → just checking exceptions
beforeEvery month-end, HR "disappeared" for two full days — tallying attendance, calculating deductions, building pay statements.
nowAI drafts the tallies and statements; a person only checks the exceptions and approves. That "two days at month-end" is gone.
What we hand AI is the manual work of tallying and drafting statements. We don't entrust payroll itself to it (the line on what's okay to enter is in #5 and #7).
04 PRDrafting short videos
Prep work → focusing on the work of taste
beforeOne draft took a whole morning — organizing footage, picking cuts, drafting captions, laying out the timeline.
nowFrom organizing footage to drafting the timeline, AI handles it. Selection and final editing — the work of taste — stay with a person.
05 Company-wide "Tachikoma"New-hire support bot
Ask a senior → 24/7, self-sustaining
beforeNew hires asked "Sorry — how do I do this?" many times a day. Seniors' focus was chopped up every single day.
nowThe Slack bot "Tachikoma" reads the internal wiki and past threads and answers instantly. Seniors step in only when it's hard. New hires can ask anytime, 24/7, and keep going on their own.
We grow "can do it" until it stays in the company.
From getting everyone on the same page, to using it in your own work, to hands-on coaching, to embedding it in the workflow — shaped so it's still used outside the training room.
Get on the same page
From leadership to the front line, the basics of what AI can and can't safely do, in shared words.
Try it on your own work
Not just lecture — right there, let AI take on your actual current task.
In our case, a new sales hire we handed AI to started making AI do their own prospecting work right on the spot.
Hands-on coaching
Beside you where you get stuck. We hand over patterns for how to ask AI, and examples that work.
Drawing the safety line
Which information is okay and which isn't, with concrete examples. In-house rules you can use with confidence, set together.
Embed it in the work
Place AI inside the flow of daily work, shaped so it's still used outside the training room.
Keep using and spread it
Toward a state where those who can do it can teach the next person. It stays as a way of working inside the company.
By now the new-hire support bot "Tachikoma" has grown to where new hires can keep going on their own, 24/7.
We don't end it in a single day of training. We stay with it until it takes a shape you can keep using.
"Too soon for us" and "company-wide is hard" — we clear both up front.
You don't have to go company-wide at once. You can start small, from one team and one task.
It doesn't have to be company-wide
Start from one team, one task. If it doesn't fit, you can stop.
Gaps are natural
Assuming a mix of confident and unsure people, we coach each of them accordingly.
Only the information that's okay to enter
We walk through the line between what's okay and what isn't, with concrete examples, right at the start.
Training only, or combined
Together with an assessment or development, or training alone — either is fine. Adding more later is up to you.
If you'd rather decide where to start first, see "AI Consulting (assessment)." If you already know what you want built, see "Custom Development." Training-only inquiries are welcome too.
First, 60 minutes to look at where it isn't working yet, together.
"We don't know what to teach" is a perfectly fine starting point. Together we sort out which team's and which task's start would stick best, and hand it to you. For a sales team, say, "build your own prospect list"; for accounting, "auto-collect receipts" — we propose from a concrete first step like that.
Free AI assessment — what you take away in 60 minutes
- A read on whether training is needed
- Which team and which task to start with
- A proposal for the training content and how to proceed
Common, honest questions
Can we request just the training?
Yes. Training alone is fine. You can also combine it with an assessment or development, but starting with training, or training only, is very welcome. You can begin from the one team or one task that's giving you the most trouble right now.
How is this different from "AI Consulting"?
AI Consulting is the service where your company decides what to start with and where to apply AI. Generative AI training builds what comes next — employees actually becoming able to use it, and keeping it up. Think of it as consulting when you want to decide the direction, and training when you want to grow the number of people who can use it and make it stick. Of course, you can also ask about both together.
How much does it cost?
The entry-point free AI assessment is 0 yen (60 minutes). The cost of the training itself varies with the number of people, the number of teams, and how far the coaching goes, so we don't publish a flat price. After hearing your situation in the assessment, we provide a quote you can take through internal approval, starting from a small, comfortable scope. It never jumps to a large contract.
How many people, and in what format?
One team of a few people is fine. Rather than company-wide at once, you can start small from one department and one task. Group sessions, small groups, or online — we fit your preferred format. If it doesn't suit you, you're free to stop partway.
What exactly do you teach?
The basics of what AI can and can't safely do, patterns for how to ask AI in your own work, examples that work, and the line between information that's okay to enter and information that isn't. Rather than lecture alone, we go as far as letting AI take on your actual current task right there. The content is built to fit your company's work.
Can AI be used in back-office or creative work too?
Yes. We ourselves run everything from logging accounting receipts to drafting PR videos to HR's month-end work with AI every day — back-office and creative alike. It isn't a tool just for engineers. If anything, the more repetitive manual work a job has, the more parts you can hand to AI. In the training, we explore "what can be handed over" together, fitted to that person's actual work.
Won't it fail to stick anyway, even with training?
That concern is entirely reasonable. That's why we don't end at a single day of training — we go together through trying it on your own work, hands-on coaching, and embedding it into the flow of work. We won't say "it will definitely stick" — we stay with it as far as shaping it so it's likely to stick. If it doesn't suit you once you try, you can stop.
What about leaks and safety rules?
In line with the AI Business Operator Guidelines from Japan's Ministry of Internal Affairs and Communications and the Ministry of Economy, Trade and Industry, we set up the rules for safe use together — the line between information that's okay and not okay to enter, records of who used it and when, and the steps for human review. We walk through it with concrete examples during the training.
So it doesn't end at "learned it and done."
Start with a free 60-minute assessment, and together we'll sketch the path to employees being able to use AI.