Decide · Design · Ship
How does one team take an idea to a live product without losing its point? In three steps, run by the same people: decide what's worth building and what gets cut, design the screens people will actually use, and ship it tested, staged and observable.
Three steps, one team, no handovers.
What moves it to the next step
Each step hands the next one something written down, and the same people pick it up, so nothing has to be explained twice.
Decide
What's worth building, for whom, and what gets cut before any code is written. This is where most of the money is saved: every feature cut here is one nobody has to design, build, test or maintain. The Scope Cut sorts the ideas, and Kill Criteria set the result that would make you stop.
A scope in writing: the buyer, the first release, what was cut and why, and the stop line, agreed by the person who decides.
Design
The screens people actually use, designed for how they already work rather than how the product would like them to. For an AI product that includes where the model's answers appear, and what happens when it's wrong or unsure.
The people who'll use it can do the core job on the design, in the order they really do it, without being walked through it.
Ship
Built on real data, tested like any other code (the model included), released in steps, and watched once it's live. The code and data sit in your accounts from the first day, not ours.
It passes its tests on real data, goes out through a graduated launch, and every step it takes can be traced.
apprn: agentic salon management.
Own venture apprn, run end to end by the same team: thesis, brand, product, build.
Decide
In a salon, money leaks where nothing can be checked. We decided the product's spine would be one verified event, not a model: an artist can only start a service by entering the one-time code sent to the customer with that booking. AI goes only where judgement helps.
Design
For people who don't think of themselves as software users. Customers stay on WhatsApp, with nothing to install; artists get their own app; and pay follows the record, visible to owner and artist alike.
Ship
A live test with real salons. The go or no-go decision is made against kill criteria set before the test, and we'll publish the criteria before the result.
Where it goes wrong
- Misread
Three steps, three teams
Why Every handover loses the reason a decision was made. The point of the sequence is that the people who cut a feature are the people who design around its absence.
- Misread
Designing before deciding
Why Screens make an undecided scope look finished, and a finished-looking feature is much harder to cut.
- Misread
Treating ship as a date
Why Shipping includes the tests, the staged release and the means to see what the product is doing. A launch without them moves the hard part to the day after.
In the Journal: Define: what exactly are we building? and Build: does it exist, and does it work?
Published 27 September 2026. Graylemon original, the studio's method.