Coding with AI: A Practical Approach
Welcome to my guide on effectively using AI for coding. Based on extensive experience building numerous projects over the past year, I have gathered practical and easily implemented steps on how to maximise AI tools for development while avoiding common pitfalls.
I also touch on the use of CodeGuidedev, a useful tool that allows you to create comprehensive documents for creating projects. I have also added a comprehensive list of the types of documents they offer.

by Reita Williams

The Reality of AI Coding
Daily Practice
I code with AI tools every day, integrating them into my regular workflow.
Extensive Experience
Built numerous projects in the last 12 months using AI assistance.
Hard Truth
The "Build me [X] app" approach in a single prompt simply doesn't work.
While AI coding tools do help and accelerate development, they aren't magic. Success requires understanding their limitations and learning how to work with them effectively. Let's explore the common mistakes that prevent developers from harnessing AI's full potential.
Mistake 1: No Planning
Voice Conversation
Turn on voice with your preferred chatbot and have a one-on-one conversation about your project goals.
15-Minute Chat
Brainstorm ideas, get critique, and conduct web research through conversation.
Draft Creation
Ask AI to "write me a well-structured draft on all the things we've DONE."
MVP Features
End with a one-page document outlining core features of your MVP.
Don't build blindly. Using AI as a planning partner helps clarify your vision before writing a single line of code. This approach transforms vague ideas into actionable development plans in just hours.
Mistake 2: No Knowledge Base for AI

Complete Documentation
Comprehensive reference for AI
Technical Guidelines
Frontend & Backend structure
App Architecture
PRD & App flow documents
After drafting your idea, create a knowledge base for AI coding models using @CodeGuidedev by writing {coding docs}. This crucial step significantly reduces AI hallucinations during development.
Your documentation should include a Product Requirements Document (PRD), app flow documentation, tech stack specifications, frontend guidelines, and backend structure. These documents serve as reference points that AI can consult to understand what to build next and what to avoid.
Mistake #3: Not Picking the Right Tools
Every AI development tool excels at specific tasks. Choosing the right tool for each job dramatically improves your results:
For Beginners
Windsurf AI offers the best IDE experience for those starting out
Full Stack Development
Cursor AI provides excellent support for comprehensive applications
Documentation
CodeGuidedev specialises in creating project coding documentation
Rapid Prototyping
Replit excels at quickly creating MVPs and mobile applications

Tool Breakdown

@windsurf_ai - is best IDE for beginners (1st right now) @cursor_ai - is great for full stack apps (2nd right now) @CodeGuidedev - is great for writing project coding docs @Replit - is great for one-shotting MVPs/mobile apps @v0 - is great at designing modern UI components @boltdotnew - is great for Micro SaaS using Supabase @lovable_dev - is great for coding modern landing pages @SoftgenAI - is great for Micro SaaS with firebase @cline - is best VS Code extension/works with MCPs @base_44 - is great for dashboard-like apps (new tool) @ManusAI_HQ - is great for agentic/browser based work @perplexity_ai - is best for web search/summarization @ChatGPTapp - is great for voice to voice/writing work @grok - 3 is great at web based research/X information @AnthropicAI - is great for Coding (Sonnet 3.7 model) @OpenAI - is great for all types of AI models via API @GeminiApp - is great for multimodal (Flash 2.0 model)

Mistake #4: Wrong Tech Stack
Frontend
  • NextJS
  • Vite
  • Flask
Database
  • Supabase (PostgreSQL)
  • Firebase
Authentication
  • ClerkDev
  • Supabase
  • Firebase
AI Integration
  • OpenAI
  • Claude
  • Gemini
AI models are trained on specific coding languages and frameworks. For optimal code quality and fewer errors, stick to AI-friendly tech stacks. These technologies have extensive representation in training data, allowing AI to generate more reliable and efficient code.
Mistake 5: Not Building Step by Step
Create Detailed Plan
Develop a comprehensive implementation roadmap before coding
Follow Structured Approach
Use tools like CodeGuidedev's 50-step implementation plan
Execute Methodically
Use AI only to implement specific steps, not to plan them
When you allow AI to plan the next steps in development, it will mess up the codebase 8 out of 10 times. Instead, use AI models solely to execute your predetermined plan and implement the code according to your specifications.
A detailed, step-by-step approach forces AI to follow your vision precisely, preventing it from missing critical components or introducing unwanted features.
Mistake 6: Poor Debugging Approach
Chain of Thought Reasoning
Ask AI to "use chain of thought reasoning to find the core issue first and then plan step by step to fix the issue."
Best Practices
Request that AI "follow the best practices of code" when proposing solutions.
Web Search Integration
Instruct AI to "search the web and find the fix for this issue" for challenging problems.
Focused Context
Only attach relevant files so AI can focus better on the specific problem.
Debugging is often the most frustrating aspect of AI coding. To make this process less painful, use specific prompting techniques that guide the AI toward methodical problem-solving rather than guesswork.
Mistake 7: Using Only One AI Model
Claude Sonnet 3.5
Best for executing code in Cursor/Windsurf (better than 3.7 for this specific task)
GPT o1/o3-mini-high
Excellent for debugging complex errors and finding solutions
Gemini Flash 2.0
Ideal for scanning complete codebases and updating documentation
A single AI model cannot excel at every development task. Different models have different strengths, and leveraging multiple specialised models significantly improves your development workflow.
By strategically selecting the right model for each specific task, you can optimise both performance and results while minimising frustration.
Mistake 8: Not Using Starter Kits
Pre-configured Components
Starter kits come with essential components already installed and configured, eliminating repetitive setup tasks.
Accelerated Development
Starting with boilerplates dramatically reduces development time and allows you to focus on unique features.
AI-Optimised Templates
CodeGuidedev offers 6 specialised boilerplates specifically designed for AI coding models.
Starting from scratch with each project wastes tokens, time, and energy on fixing predictable errors. Using starter kits with pre-installed components allows you to build faster while maintaining consistency across projects.
Mistake 9: Quitting Too Early
Initial Excitement
AI coding is fun and exciting during the first few prompts
Frustration Phase
After the third prompt, you start encountering errors and layout issues
Debugging Struggle
Hundreds of errors, build issues, and codebase problems emerge
Persistence Payoff
With strong documentation and rules, you can successfully guide AI to completion
The initial excitement of AI coding quickly fades when you encounter inevitable challenges. Many developers abandon projects at the first sign of difficulty, missing out on the rewards of persistence.
Remember that overcoming these obstacles is part of the process. With proper documentation and a structured approach, you can successfully navigate through the frustrating middle phase to achieve impressive results.
Key Takeaways

Plan Before Coding
Create a clear vision before opening any AI tool
Document Everything
Write detailed coding docs using CodeGuidedev
Choose Wisely
Select the right AI tools and tech stack
Master Prompting
Debug effectively and use specialised models
Success with AI coding requires thoughtful preparation, proper documentation, strategic tool selection, and effective communication with AI models. By avoiding these common mistakes, you'll dramatically improve your development experience and outcomes.
Remember to use AI-friendly tech stacks, follow a step-by-step approach, and persist through challenges to achieve your development goals.
Join Me On This Journey
12+
Months Experience
Practical AI coding experience
9
Common Mistakes
Identified and solved
15+
AI Tools
Evaluated and compared
If you want to know more about effectively coding with AI, join me on my learning journey. Let's figure out this AI thing together, sharing insights and discoveries along the way.
The field of AI-assisted development is constantly evolving, and by collaborating and sharing experiences, we can all become more effective at leveraging these powerful tools to build better software faster.
© 2025 Reita Williams. All rights reserved. This content reflects personal experiences and insights. While you're welcome to use these strategies, please credit appropriately if sharing or referencing this material.
Website Quick Links
Linkedin
© 2025 Reita Williams. All rights reserved. This content reflects personal experiences and insights. While you're welcome to use these strategies, please credit appropriately if sharing or referencing this material.