I used to prompt ChatGPT all wrong
(hereโs my secret to better outputs)
Youโve heard it before
Garbage in = garbage out
Yet, must people prompt like, โwrite me a linkedIn postโย
And expect a masterpiece
Everyone wants ChatGPT to work better
But, they speak itโs language
Hereโs how to 10x your output quality
While spending less time on your prompts
Theyโre called โslash commendsโ
Think of them like ChatGPT shortcuts
There are some of my favorites
(save them to 10x your ChatGPT game)
1๏ธโฃ Formatting (Structure & Format Results)
Purpose: Make responses clear, structured, and easy to use.
/EXEC SUMMARY โ quick executive-style summary
/STEP-BY-STEP โ logical, numbered breakdown
/CHECKLIST` โ actionable boxes
/FORMAT AS`, `/TABLE-IT`, `/JSONIFY`, `/CSVIFY` โ control output formats
/SCHEMA` โ create structured outlines or frameworks
/REWRITE AS` โ restyle or refine tone
2๏ธโฃ Tone & Voice (Control Writing Style)
Purpose: Match tone and audience for your message.
/AUDIENCE โ tailor writing to execs, engineers, etc.
/TONE โ change voice (formal, funny, direct, etc.)
/ACT AS โ roleplay as CRO, PM, or writer
/BRIEFLY โ concise output
/JARGON โ technical style
/HEMINGWAY โ simplify and tighten
/BRAND VOICE โ match personal or company tone
3๏ธโฃ Analysis (Better Reasoning & Insight)
Purpose: Deepen critical thinking and analytical structure.
/ELI5 โ simplify complex topics
/SWOT โ strengths, weaknesses, opportunities, threats
/COMPARE โ side-by-side evaluation
/MULTI-PERSPECTIVE โ multiple viewpoints
/FIRST PRINCIPLES โ logic from fundamentals
/PITFALLS โ identify risks
4๏ธโฃ Quality (Improve Output & Guardrails)
Purpose: Ensure accuracy, depth, and ethical quality.
/REFLECTIVE MODE โ self-evaluate the response
/SYSTEMATIC BIAS CHECK โ detect bias
/DELIBERATE THINKING โ slow, reasoned replies
/NO AUTOPILOT โ forbid shallow answers
/EVAL-SELF โ self-grade and improve
/GUARDRAIL โ define strict tone or scope
/SOURCES ONLY โ force citations
/RUBRIC โ create grading criteria
5๏ธโฃ Research (Better Data Synthesis)
Purpose: Summarize, merge, and control information flow.
/TLDL โ quick summary of long text
/CONTEXT STACK โ keep prior context
/BEGIN WITH /END WITH โ shape narrative flow
/SYNTHESIZE โ merge multiple sources
/QUOTES-ONLY โ extract key lines
/OPEN QUESTIONS โ list whatโs missing
The math is simple:
Old way: 10-minute prompt engineering
New way: 1 slash command
While everyone else writes paragraphs...
You're getting results in seconds.
P.S. Save this. Your future self will thank you.