The Scope Cut
What goes in the first release? Every idea gets one of four verdicts: build first, later, cut, or don't build, each with a reason. The first release is what survives.
Four verdicts, each with a reason.
In: every idea on the list
- KeptBuild firstThe first release can't prove its point without it.
- ParkedLaterWorth building once something else exists first: data, usage or proof.
- TrimmedCutReduced to the part people will actually use.
- DroppedDon't buildNot needed, or something that already exists does the job.
Out: the first release, and only what it needs
What each verdict needs
Build first
The first release exists to prove one thing. An idea is built first only if that proof fails without it.
You can say what the first release can't show without it.
Later
Worth building, but it depends on something that doesn't exist yet.
The reason names what has to exist first, so the idea comes back when it does.
Cut
The idea stays, smaller. Only the part people will use survives.
You know which part gets used, from watching people work, not from asking what they'd like.
Don't build
Either nobody needs it, or something that already exists does the job well enough.
The existing option is named, or the missing need is.
Six ideas, two in the first release.
Illustrative example A clinic booking assistant. Not client work.
| Idea | Verdict | Reason |
|---|---|---|
| Booking over WhatsApp, with reminders | Build first | Customers are already there. |
| Payment at the time of booking | Build first | It cuts no-shows with no AI at all. |
| A custom-trained model | Don't build | An off-the-shelf model does this job. |
| An admin dashboard with 14 reports | Cut to 2 | Nobody reads twelve of them. |
| No-show prediction | Later | It needs three months of real data first. |
| A voice assistant | Later | Not until the text flow is proven. |
The first release is two ideas built as one flow, in the chat customers already use. It's the same example our homepage walks through, stage by stage.
Where it goes wrong
- Misread
Cutting to hit a date
Why The cut decides what proves the idea. Scope is fixed first, and the end date is set for that scope.
- Misread
Treating Later as a polite no
Why Every Later names what has to exist first, so it comes back when that exists.
- Misread
Keeping everything, smaller
Why A first release with every feature at half quality proves nothing about any of them.
- Misread
Verdicts without reasons
Why A verdict without a reason gets reopened in the next meeting.
Download: the AI product scope template, a free Word file.
Try it in the Lab: Cut It sorts a real brief into build, later and cut, on a fixed budget.
In the Journal: Is it worth building? · White-label AI for agencies
Published 24 September 2026. Graylemon original, and the demonstration on our homepage.