2 Prompt Mastery – Why Less Is More (7 mins)
Module 2: Keep It Simple – Why Minimal Custom Instructions Work Best
Intro:
Setting up custom instructions in AI tools like ChatGPT and Claude can personalise your experience, but too much detail can make AI responses rigid or inconsistent. This module covers how to optimise your custom instructions for flexibility and better results.
Learning Objectives:
- Understand what custom instructions do and how they impact AI responses
- Learn why minimal instructions work better than overly detailed ones
- Discover where to store context-specific details instead of top-level settings
- Practise refining your own AI instructions for better adaptability
written version of the video tutorial
AI tools like ChatGPT and Claude allow you to set custom instructions to personalise how they respond to you. While this might seem like a great way to fine-tune AI for your needs, too much detail can actually reduce AI’s effectiveness.
In this module, you’ll learn:
- What custom instructions are and how they work
- Why keeping them minimal leads to better AI interactions
- The difference between top-level instructions and context-specific prompts
- Where to store more detailed instructions for flexibility
- A simple exercise to refine your AI’s custom settings
By the end of this lesson, you’ll have an optimised setup that makes AI work better for you, not against you.
What Are Custom Instructions?
Custom instructions allow you to set default preferences for how AI responds to you.
For example, in ChatGPT, you can customise:
✅ What it should know about you to provide better responses
✅ How you’d like it to respond
This means AI can tailor its tone, detail level, and style to match your preferences.
While this sounds useful, many users make the mistake of overloading custom instructions with too much information. Instead of improving AI’s responses, this can make them rigid, unnatural, or inconsistent.
Why Too Much Detail Can Backfire
If you add too many constraints in your custom instructions, you may experience:
❌ Rigid Responses – AI forces all answers into a narrow format, losing adaptability.
❌ Conflicting Priorities – If you ask for concise responses but also highly detailed explanations, AI may struggle to balance the two.
❌ Unnatural-Sounding Output – AI might overcompensate for instructions, leading to awkward or robotic responses.
For example:
💥 Overloaded Custom Instruction:
“Always write in a highly detailed, formal, and structured way with long paragraphs. Avoid humour, contractions, and casual language at all times.”
🔴 Problem: This forces AI into one strict style, making it less useful when you need a quick or conversational response.
✅ Minimalist Custom Instruction:
“Use British English and provide clear, structured responses in a professional yet conversational tone.”
✅ Benefit: This keeps AI flexible while maintaining your core preferences.
The Best Approach: Minimal Top-Level Instructions
A better approach is to only include instructions that apply across all conversations and store context-specific details elsewhere.
What Should Go in Top-Level Custom Instructions?
✅ Who you are – “I work in education and create online learning materials.”
✅ Preferred language and style – “Use British English and avoid corporate buzzwords like ‘delve’ or ‘crucial’.”
✅ Basic tone preferences – “Provide clear, structured responses in a professional but natural style.”
Anything that only applies in specific scenarios should be stored elsewhere.
Where to Store Detailed Instructions Instead
For task-specific AI guidance, store instructions inside the right tools, not the top-level settings.
📌 Use Custom GPTs (ChatGPT Plus Users) – Set up separate AI assistants for different workflows.
💾 Save Prompts in a Library (Google Docs, Notion) – Store detailed prompts for quick access.
⭐ Use Starred Chats (Claude Users) – Save useful AI interactions for future reference.
Example: Using a Saved Prompt Instead of Overloading AI
❌ Bad Approach (Overloading Top-Level Instructions):
“Always write in a humorous tone and include jokes in every response.”
✅ Better Approach (Saved Prompt Instead):
Create a separate prompt called ‘Humorous Writing Assistant’ that you use when needed.
This way, AI remains flexible across different tasks.
Practical Exercise – Refining Your Custom Instructions
Time to put this into action!
1️⃣ Open your AI’s custom instructions (ChatGPT, Claude, or another tool).
2️⃣ Review what you’ve written – Have you added too many details that don’t apply across all interactions?
3️⃣ Simplify your instructions so they only include:
- Who you are
- Basic language and style preferences
4️⃣ Move detailed, task-specific instructions into your prompt library or AI tool settings.
5️⃣ Test it out! Ask AI a question and see if the response feels more natural.
📌 Bonus Tip: If you’re unsure, try removing half of your current instructions and compare the AI’s responses before and after.
Summary & Next Steps
Key Takeaways:
✔️ Custom instructions should be minimal to allow AI to remain flexible.
✔️ Overloading AI with too many constraints can lead to rigid, unnatural responses.
✔️ Only include top-level preferences like language, tone, and evergreen details.
✔️ Store detailed, task-specific instructions in saved prompts, custom GPTs, or starred chats.
📌 Action Step: Review and refine your AI’s top-level instructions today!
🚀 Next Module: Now that you’ve optimised your AI’s setup, we’ll explore how to experiment and iterate prompts for even better results. Let’s dive in!