Why a Writer
Builds with AI.
Every project at biw.studio is a two-person relay between a human and AI.
I build the concept, choose the words, and carry the final responsibility.
The speed of writing code, and the thoroughness that catches what I'd miss — that's AI.
It's not something to hide, or to be embarrassed about. It's the way of working I chose, so I'm telling you about it first.
Q1 What do I hand off to AI?
Listening to what a client needs. Deciding what to sell. Designing the words. Inspecting the work before delivery. — That's all human work, mine.
Turning a finished design into something fast and precise. — That's AI's job.
Of the five steps in my Story Design Method, AI only takes the lead in one: "the hand that orders, the hand that guides" — the step where the finished design gets built. The detailed division of labor is on the study page (Japanese only).
Q2 Why do I use AI?
Three reasons.
First: it's fast. The time from consultation to a working draft shrinks, and I can turn around edits the same day.
Second: the finish is more consistent. AI checks every corner with eyes that never tire, and then I confirm, as the human, whether it's correct — and whether it feels right. Two sets of eyes remove the unevenness.
Third: you get to compare and choose. Being able to put two design drafts side by side and let you pick by feel is only possible because I work with AI.
And the time that gets freed up by working faster doesn't turn into a discount — I pour it back into concept work and inspection. That's what actually shows up in the result.
Q3 Who's responsible for quality?
I, the maker, am the final person responsible for everything biw.studio delivers.
Every deliverable passes through my own eyes before it ships: checking facts, checking copyright, checking it against what you asked for.
"AI wrote it" is never an excuse. If there's a mistake, it's my mistake — I fix it, and I own it.
Q4 Is our information safe?
Your confidential information (customer lists, sales figures, personal data, and the like) never goes into AI.
Information I do need for the work (your business details, product descriptions, and so on) is handled with model training turned off.
Exactly how far AI's reach goes is always explained before the work starts, and I only work within the scope we agree on. I always sign an NDA, too.
Q5 Which AI do I use?
Mainly Claude, the AI assistant from Anthropic, a company known for its AI safety research.
I chose it for its strength with long documents and code, and for how much control it gives me over how my data is handled.
Which AI I use shifts by step. I also run finished work past a separate AI for review — keeping the eyes that build separate from the eyes that check.
The tools will change as they improve. What won't change is that I name them, right here, before you have to ask.
Four Promises
One. The final call and the inspection are always human.
I confirm facts, copyright, and fit with your request with my own eyes before delivery.
Two. Your confidential information never goes into AI.
What AI touches is agreed in advance, and I run it with model training turned off.
Three. Responsibility never sits with AI.
I'm the one accountable. Mistakes and bugs alike, I fix as my own.
Four. I don't hide the tools I use.
I publish them here, before you have to ask.
One Last Thing
I no longer take on projects without AI. Not to cut corners — this way of working is simply the best work I can offer right now.
The research, the data, and the reasoning behind this policy are on the study page (Japanese only), with sources cited.