←Back/Product/Codex
IntermediateProduct

Build a Native Mobile App with AI and Clear Product Direction

Use Codex to help build a native mobile app, then guide the design and review the resulting code. William brought web-development and product experience to the project.

Illustration for How to Build a Native Mobile App with AI, Even If You Can't Code
Source episode · 00:09:02William Imoh Built a Norwegian Tutor That Refuses to Speak English

with Claire Vo

I first started with a very long prompt that was very specific, and this prompt opened with literally, 'I want a very aesthetically pleasing application.' ... I described the experience I want... I wanted an orb that was, you know, a pulsating orb that I could talk to... I wanted something warm. I literally wrote, 'I want something that feels like fall.'

Read the full episodeWatch on YouTube

Before you start

What you need

  • A clear product concept and user experience goal.
  • Specific design vocabulary and visual references.
  • Product and engineering experience, or access to an experienced reviewer for architecture, security and testing.

What you’ll make

The generated source code for a native mobile application that aligns with the specified design and functionality.

Tools used

  • Codex

    OpenAI's cloud-based AI software engineering agent that can execute code, run tests, and handle complex multi-file tasks autonomously.

    VisitCodex

The workflow

Follow the sequence once, then adapt the prompts, checks, and handoffs to your own setup.

4 steps

Step 01

Write a Detailed Kickoff Prompt

Begin by writing a comprehensive prompt that describes the application's purpose, target user, and desired experience. This is the most critical step, as it sets the foundation for the entire project.

Example prompt
I want to build an aesthetically pleasing [iOS/Android] application for [purpose, e.g., language learning]. The main screen should feature a [central UI element, e.g., pulsating orb]. The color palette should feel [adjective, e.g., warm, like fall]. The interface must be [style, e.g., minimalist], with no unnecessary elements. Follow the design best practices of a [brand reference, e.g., senior designer at Apple].

Structure your prompt like a product epic or user story. The more specific you are about the experience, technology stack (including a currently documented voice API), and aesthetics, the better the initial result will be.

Step 02

Inspect and Critique the First Version

Ask the AI to generate an initial version of the app. Your job is to act as the product owner and designer, identifying what doesn't align with your vision.

Don't expect perfection on the first try. William's first result was 'shit' because it was in dark mode, which didn't match the 'warm' feeling he wanted. Be prepared to reject the initial output and provide clear reasons why.

Step 03

Refine with Specific Design Vocabulary

Provide corrective feedback to the AI using precise language. This is where having a design vocabulary is more important than knowing how to code.

Example prompt
The current design is in dark mode. Change it to light mode. Use light, warm colors. Explore using gradients and a liquid glass effect for the main UI elements.

Iterate on this step. The conversation between you and the AI refines the product. Test each revision against your requirements.

Step 04

Request Core Functionality and Constraints

Once the visual design is closer to your goal, prompt the AI to implement the core logic and any unique behavioral rules that define your app.

Example prompt
Implement the primary function, which is a voice conversation with the user. The app must use [specific technology, e.g., OpenAI's realtime voice model]. A key rule is that the app must [behavioral constraint, e.g., refuse to switch languages when asked].

Clearly define what the app should and, just as importantly, should not do. These constraints are often what make the product valuable.

What good looks like

  • The AI generates functional code in the desired native language (Swift, Kotlin).
  • The initial app version reflects the core concepts of the kickoff prompt.
  • The AI successfully incorporates specific design feedback in subsequent iterations.

Build your next product with ChatPRD

Turn an idea into a PRD, user stories, and a plan.

Try ChatPRD free

Runbook notes

How to recover when the loop fails and where human judgment helps.

If it goes sideways

The AI produces a generic or aesthetically poor design.
Refine your prompts with more specific design language. Instead of 'make it look good,' use concrete terms like 'use light mode, not dark mode,' 'add gradients,' and 'use a liquid glass effect.' Reference the style of a well-known brand, like 'follow design best practices like you’re a senior designer at Apple.'
The AI gets stuck, produces buggy code, or misunderstands a request.
First, try restarting the agent and repeating the prompt. If that fails, become more explicit: break the task down and provide step-by-step instructions like 'do this first, test this first, then review what you've done.'

Human checkpoints

Let routine work keep moving.

  1. Review the initial design generated by the AI to ensure it matches your vision before proceeding with feature development.
  2. Approve the final user interface and experience before considering the application complete.

Related workflows

Illustration for Create an Immersive AI Language Tutor
intermediateProductCodex

Create an Immersive AI Language Tutor

Build a voice-first language practice app that forces immersion by refusing to switch to the user's native language. The AI tutor provides corrections in real-time, within the flow of the conversation, to simulate an authentic learning environment.

Open
Illustration for How to Manage Business Operations with Role-Based AI Agents
advancedOperationsChatGPTCodex+1

How to Manage Business Operations with Role-Based AI Agents

Organize your business operations by creating a 'fellowship' of specialized AI agents. This workflow shows how to silo responsibilities into projects and assign distinct roles to AI assistants to automate administrative tasks and context retrieval.

Open
Illustration of Build a Personal PM Operating System with a Shared Knowledge Repo
intermediateProductCodexClaude

Build a Personal PM Operating System with a Shared Knowledge Repo

Create a repository-based system to provide an AI assistant with durable context about your team, projects, and communication channels. This enables it to perform routine tasks like summarizing daily updates and identifying stakeholders, acting as a personalized chief of staff.

Open

Start shipping better products.

Join 250,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.

Start building freeRequest a demo
Free to start · 14-day Pro trial · SOC 2 Type II