Imagine.bo

Best Practices

Get the most out of every prompt and build faster, smarter apps from day one.

Write Clear Intentional Prompts

Vague prompts produce generic results. The more context you give Imagine.bo upfront, the more accurate and complete your generated app will be.

Every great prompt contains four core elements:

  • Actor / Persona — Who is this app for? (e.g. "small business owners", "project managers", "customers booking appointments")
  • Problem to Solve — What core challenge does the app address?
  • Key Features — List what the app must do explicitly.
  • Edge Behaviors — Include rules, exceptions, and constraints (e.g. "Only admins can delete records", "Tasks overdue should notify the assigned user").
Good example

Build a customer support ticketing app for small businesses. Users can submit tickets with a priority level, upload attachments, and receive email notifications. Admins can assign tickets to agents and manage ticket status.

Poor example

Build something like Jira.

Start with MVP Scope

Trying to build everything at once makes the first generation slow, noisy, and harder to refine. Start with the minimum viable feature set:

  • Prioritize the core problem only.
  • Include only the essential screens and user flows.
  • Avoid loading one prompt with a Version 1 and a Version 2 feature list.

You can always add advanced features after your MVP is live.

Define User Roles and Access Early

Access control errors are the most common source of confusion in generated apps. Prevent data leaks and permission issues by specifying roles clearly in your initial prompt.

  • State which roles exist (e.g. Guest, Member, Admin).
  • Define what each role can and cannot do.
  • Restrict sensitive data by role explicitly.
Example
Admins can approve or reject requests. Only logged-in users can create posts. Guests can only view public listings.

Use Conversation to Refine Not Rebuild

Imagine.bo's AI understands follow-up instructions. Instead of starting a new project when something is off, simply continue the conversation:

  • Change layouts: Move the analytics chart to the top and make it full-width.
  • Update logic: Require email verification before a user can post content.
  • Adjust design: Make the header dark mode and use a sans-serif font throughout.

Iterating through conversation is faster and preserves everything already built.

Test Workflows Before Deploying

Before you hit Deploy, run through your key user journeys manually:

  • Test with multiple user roles (admin, standard user, guest).
  • Check edge cases: What happens with invalid input? What if a required field is empty?
  • Confirm that permissions are enforced correctly at every step.

Use the Hire a Human Feature Strategically

The AI is powerful, but some tasks benefit from human engineering. Use the Assign a Developer feature when you encounter:

  • Complex third-party API integrations (e.g. localized payment gateways).
  • Custom algorithms or data processing logic.
  • Performance tuning and optimization.
  • Security-sensitive features that require human review.

Frame your developer task clearly — include the goal, the expected result, and any reference examples.

FAQ

Actor or Persona who the app is for, Problem to Solve the core challenge, Key Features what the app must do explicitly, and Edge Behaviors rules exceptions and constraints

MVP stands for Minimum Viable Product — starting with the minimum viable feature set makes your first generation faster and easier to refine, you can always add advanced features after your MVP is live

State which roles exist such as Guest Member Admin, define what each role can and cannot do, and restrict sensitive data by role explicitly in your initial prompt

Use follow-up conversation prompts to change layouts update logic or adjust design, the AI understands follow-up instructions and iterating through conversation is faster and preserves everything already built

Use Assign a Developer for complex third-party API integrations, custom algorithms, performance tuning, or security-sensitive features that require human engineering review