# A feature brief for a useful first change

By Abdullah Faruk Gönüllü · afaruk.dev

Use this with your developer. Leave unknowns visible and assign the decision to a person.

## The user and the job
- Who is trying to do what, in which situation?
- What happens today? Include one real example.
- Why is this worth improving now?
- How will we tell whether the user's job improved?

## The first useful scope
- What is the smallest complete behavior we can release and learn from?
- What is explicitly excluded?
- Which records are authoritative, and who maintains them?
- What permissions apply?
- Which unanswered decisions must be resolved before implementation?

## Acceptance scenarios
- Given [situation], when [action], then [observable result].
- Include the important denied-access, missing-data, and failure cases.
- State what existing behavior must continue to work.

## AI assistance and ownership
- What may AI draft, change, or execute?
- Which codebase conventions and checks apply?
- Who owns business scope, technical review, verification, and release?
- What needs a human decision before work continues?

## Evidence and release
- Reviewed change and exact version:
- Relevant checks, actual results, and environment:
- User-flow observation:
- Known limits, failures, and unresolved risks:
- Release owner and decision:
- Monitoring and support owner:
- Recovery method, including data changes:
- Post-release confirmation and user feedback:

Guide: https://afaruk.dev/blog/ai-assisted-development-nontechnical-founders
