Zero-Shot Chain of Thought is the simplest prompt engineering technique: just add 'Let's think step by step' to any prompt. Research shows this single phrase improves accuracy by 10-40% on reasoning tasks.
Take any prompt you'd normally give to AI
Add 'Let's think step by step' at the end
The AI will show its reasoning before giving the final answer
The reasoning process itself improves the accuracy of the final answer
Template
Example Output
Template
Example Output
This is the highest ROI prompt technique — 5 extra words for dramatically better output
Works best on: math, logic, analysis, strategy, and multi-step problems
Combine with specific step instructions for even better results
If the AI's step-by-step reasoning reveals an error, point it out and ask it to redo that step
For creative tasks, use 'Let's brainstorm step by step' instead
The phrase triggers the AI to decompose problems rather than jumping to answers. Each reasoning step creates context that improves subsequent steps. It's like thinking out loud — the process itself produces better conclusions.
Yes, it works with ChatGPT, Claude, Gemini, and virtually every large language model. It's been extensively studied and consistently shows improvement across different models and tasks.
Simple factual questions ('What year was Shakespeare born?'), creative brainstorming where you want spontaneity, and very short responses. For these, the extra reasoning steps add overhead without improving quality.
Take our free AI course and learn techniques that go beyond templates.
Start Free AI Course →