By VibeStudio, the UK technology studio led by Howard Scott and Spencer Gallagher.
What should a project brief include?
Include the business outcome, intended users, current tools and data, essential workflows, budget, deadline and acceptance checks. A short written brief is enough to begin. VibeStudio uses it to discuss fit, scope and a useful first prototype.
Name the problem and the outcome
Explain what is difficult today and what should change. Use a measurable outcome such as fewer hours spent preparing a report, a clearer enquiry process or one place to track client work. Record a baseline before expecting an improvement.
Describe the people and their main task
List who will use the result, what they need to do and who approves important actions. A public website, a staff dashboard and a client portal need different access and interaction choices.
Identify the information and tools involved
Name existing systems and give redacted examples of the information the project uses. Note who owns the data and whether an API or export is available. Do not include passwords, access tokens or private customer records in an initial brief.
Choose the smallest useful release
Separate essential tasks from later ideas. For a website this might be service information and an enquiry form; for an application, one complete workflow; for an AI agent, one bounded job with human approval.
Set acceptance and safety checks
Describe how you will know the result works. Include mobile use, accessibility, permissions, error handling and data recovery where relevant. For AI output, specify what a person must review and what should happen when the system is uncertain.
Agree the budget, timing and ownership
State any fixed launch date and why it matters. Distinguish prototype feedback from production testing, and discuss hosting, model usage, maintenance and handover. A prototype demonstrates direction; it is not proof that a system is ready for every production risk.
A brief you can copy
- Our business and the people who will use this:
- The task or problem we want to improve:
- The outcome we will measure:
- Our existing tools and the information involved:
- The essential first-release features:
- Privacy, permissions and approval requirements:
- Budget, deadline and who will maintain the result:
Which kind of build fits?
A website helps visitors understand your business and enquire. A business application supports a repeatable operational workflow. AI automation helps process information across that workflow, with human oversight where needed. Compare the service options, pricing basis and delivery process before choosing.
Technology follows the requirements. Static HTML can suit public information, a database-backed application can suit shared records, and an API integration can connect existing tools. Replit development and bounded AI agents are options, not requirements for every project.
What does real delivery evidence look like?
Read the 72-hour website and CMS case study and the HypnoBarn case study, including a client's report of around 300% more bookings in the first month. These describe individual projects, not guaranteed results or a universal delivery deadline.