Write the brief before you pick the builder.
Before hiring a builder or prompting an AI app builder, write the job, boundaries, exceptions, success evidence and handover expectations. A brief outline you can use today.

Before you choose a builder—or paste an idea into an AI app builder—produce a structured brief: the job, its boundaries, the exceptions, what success looks like in evidence and what handover you expect. That brief is the buying instrument. Tool choice comes after the work is legible.
The brief is the first product
Before you hire a builder, and before you paste an idea into an app builder, write down the job. Not the stack. Not the mood. The job: who does it today, what done looks like, what must never happen, and how you will know.
A builder, human or software, will fill every blank you leave. Some of those fills will be reasonable. Some will become the product you have to live with. The brief is how you stay in the conversation.
Fill it in this order
The job. One sentence a person outside the project can repeat. Verb and object. "Route incomplete seller packets to a named reviewer" is a job. "Transform operations with AI" is not.
The boundary. What is in, and what is explicitly out. Name the systems it may touch and the ones it must not.
The exceptions. Three real awkward cases from the last month. Who handled them, and what they needed in order to decide.
Success, as evidence. What would be true in the system of record if this worked. Not a feeling. A state someone can open.
Handover. Who runs it after launch. What they must be able to deploy, diagnose, and change without calling the person who built it.
If a section stays empty, do not send the brief. An empty section is the part that will get invented for you.
Sequence of work
Fill the brief in this order
- Job
Actor, trigger and finished state for one observable task.
- Boundaries
Data, actions and systems the first slice must not touch.
- Exceptions
Unusual cases with owner, evidence and return path.
- Success evidence
What you will inspect on review day in the real systems.
- Handover
Access, maintenance, incidents and who approves changes.
A builder cannot clarify a product you have only described as a feeling.
Questions worth asking before anyone estimates
Who does this job when the founder is unavailable?
What was the last exception that slowed you down, and who resolved it?
If two systems disagree about status, which one do people trust?
What would make you stop a pilot rather than expand it?
What does the receiving team need on day one that is not source code?
These are not a script. They are the questions that turn a pitch into an operating description. Ask them of the people who do the work, not only of the people who want the software.
What a weak brief costs
A weak brief produces a confident estimate of the wrong thing. The demo path works. The exception still lives in an inbox. Handover is a zip file and a Slack thread.
App builders fail the same way, faster. They are good at a contained workflow your team can maintain. They are a poor place to discover, halfway through, that the job needed permissions, recovery, and a source of truth.
Compare builders on how they talk about the exception and the handover. Not on how quickly they can show a screen.
Decision guide
Brief-first vs prompt-first
| Consideration | Brief-first | Prompt-first |
|---|---|---|
| Starting point | Job, boundaries and exceptions written down. | A vibe, screenshot or single prompt. |
| Supplier talk | Candidates walk your task and one failure. | Demos optimize for looking finished. |
| First slice | One actor, one finished state, agreed evidence. | Screens accumulate without a test of the job. |
| Handover | Ownership named before launch pressure. | Access and runbooks appear in the last week. |
Use the brief before you prompt or hire
If you are about to paste an idea into a builder, paste the brief first and fill the gaps. If you are about to hire, send the brief with the request. Ask each candidate to walk one ordinary task and one exception in your language.
The brief is the buying instrument. Tool choice comes after the work is legible. When you want that turned into a product, the work sits under web applications.
Related work: Sutton Place.
We build alongside the people doing the work.
Bring us the problem you’re thinking about.